A Mobile Approach to Farmer-Computer Interaction

FACULDADE DE ENGENHARIA DA UNIVERSIDADE DO PORTO A Mobile approach for Farmer-Computer Interaction Maria João Ribeiro Marques Mestrado Integrado em Engenharia Informática e Computação Supervisor: Miguel Pimenta Monteiro Co Supervisor: Bruno Giesteira Co Supervisor: Eduardo Pereira July 23, 2017 A Mobile approach for Farmer-Computer Interaction Maria João Ribeiro Marques Mestrado Integrado em Engenharia Informática e Computação July 23, 2017 Abstract Agriculture has been responsible of sustaining and enhancing human life starting at least 105,000 years ago. Nowadays, traditional approaches to farming can be enhanced with technology, helping farmers make more informed decisions when it comes to their farming activities. Information and communication technologies, specifically, mobile phones, are now enabling farmers to improve efficiency and optimize production, while preserving resources. However, due to limited technological proficiency, a problem emerges. The adoption and acceptance of these solutions are dependent on the skill for handling and understanding digital media and this can be a challenge in specific demographics, where age can be a constraint and access to technology, in rural areas, be scarce. The aim of this study was to conduct a Human computer Interaction research centred around the potential end-users of a mobile application in the agricultural field. Therefore a User-centred Design methodology was followed, allowing to narrow the scope of the study, from Agriculture in general, to Prevention and Control of the Grapevine moth in Viticulture in the Douro Region of Portugal. In this process, explicit and implicit methods were used, such as Personas, Focus Groups, Individual Interviews, Field Observation and Usability Testing. Farmers and other stakeholders were involved in the all phases of the design process, from requirements gathering, analysis, design and evaluation of the prototypes created. The outcome of this study was to accommodate and support farmers’ needs and expectations while also studying how these users interact with technology. i ii Acknowledgements First and foremost I would like to thank my family and friends for all the support given during my whole academic life. Secondly I would like to thank Professor Miguel Pimenta Monteiro and Professor Bruno Giesteira for guiding me through the process of writing this dissertation. A special thanks has to be given to Eduardo Pereira and all of those at Fraunhofer AICOS who received me with open arms and always made themselves available. At last but not least, I want to thank all the winegrowers and other members of Casa do Douro that without whom none of this would have been possible and in particular Eng. Anabela Carneiro for her valuable feedback and assistance. Maria Marques iii iv “Whatever you think your life is going to be like, just know, it’s not gonna be anything like that.” Mike Mills v LIST OF FIGURES 4.19 Prioritisation of common user tasks . . . . . . . . . . . . . . . . . . . . . . . . 51 4.20PaperSketches ................................... 52 4.21AppFlow-Start .................................. 52 4.22AppFlow-Homepage ............................... 53 4.23AppFlow-Mapping ................................ 54 4.24AppFlow-Prevention ............................... 55 4.25AppFlow-Control................................. 56 4.26CraftinSketch ................................... 62 5.1 Participants during the test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 5.2 Taskcompletionrates................................ 69 5.3 Totalerrorsandassists ............................... 69 5.4 Usermetricsbytask................................. 70 (a) Usererrorsbytask.............................. 70 (b) User light assists by task . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 (c) User heavy assists by task . . . . . . . . . . . . . . . . . . . . . . . . . . 70 (d) User total assists by task . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 5.5 Totalmetricsbytask ................................ 71 (a) Totalerrorsbytask.............................. 71 (b) Total light assists by task . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 (c) Total heavy assists by task . . . . . . . . . . . . . . . . . . . . . . . . . . 71 (d) Total both assists by task . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 5.6 SUS Scores and Sub-scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 B.1 ActivitySetup.................................... 102 B.2 Categorizing by Participant 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 B.3 Categorizing by Participant 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 B.4 Categorizing by Participant 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 B.5 Activityconclusion................................. 106 C.1 Populationpyramid................................. 107 C.2 Employment rate, persons aged 20-64 . . . . . . . . . . . . . . . . . . . . . . . 108 C.3 Workingagepopulation............................... 108 C.4 IliteracyRate .................................... 109 C.5 Share of early leavers from education . . . . . . . . . . . . . . . . . . . . . . . 110 C.6 Share of persons with tertiary education . . . . . . . . . . . . . . . . . . . . . . 111 C.7 Proportion of households with broadband connections . . . . . . . . . . . . . . . 111 C.8 Proportion of people who never used the internet . . . . . . . . . . . . . . . . . 112 F.1 ActivitySetup.................................... 127 xii List of Tables 4.1 Problem and Activity Scenarios for Small winegrower . . . . . . . . . . . . . . 42 4.2 Problem and Activity Scenarios for Medium/Large winegrower . . . . . . . . . . 43 4.3 ColourScheme ................................... 58 4.4 Iconsintendedmeaning............................... 59 4.5 Othericons ..................................... 59 4.6 DesignChanges................................... 61 4.7 UsabilityTestTasks................................. 64 5.1 Participants ..................................... 67 5.2 Summaryofresults................................. 68 5.3 SUSScoresandSub-scores............................. 71 5.4 Commentaries about each user . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 5.5 Recommendations.................................. 77 C.1 LiteracyRate .................................... 109 xiii LIST OF TABLES xiv Abbreviations DDR Demarcated Douro Region GIS Geographic Information System GPS Global Positioning System GSMA Groupe Speciale Mobile Association GUI Graphical user interface HCI Human-computer interaction HTA Hierarchical Task Analysis hi-fi High Fidelity ICT Information and communication technology ICT4D Information and Communications Technologies for Development ISCE International Standard Classification of Education ISO Internation Organization for Standardization IT Information technology IxD Interaction Design low-fi Low Fidelity PA Precision Agriculture PEU Perceived ease of use PU Perceived usefulness RDD Região Demarcada Douro SSM Site-specific management SUS System Usability Scale TAM Technology Acceptance Model UCD User-centered design UI User interface UX User Experience VRT Variable Rate Technology xv Chapter 1 Introduction The application of Information and Communications Technology (ICT) in the agricultural sector opened up a whole new world of opportunity. mAgriculture and Precision Agriculture (PA) are based on using ICTs to provide information and data to assist farmers when making site-specific management (SSM) decisions. 1.1 Context The huge technological improvements in personal electronics made it possible for farmers in rural areas to access real-time information, delivered through devices such as mobile phones and tablets. Farmers are challenged when using these devices due to limited technological proficiency and the way they interact with devices that collect large amounts of data and information will have a direct impact in their farming activities. The adoption of mobile technology in the agriculture sector is dependent of the end-users technological proficiency, so it is necessary to study the existing scenario on the general use of ICTs on the field by farmers, in order to guarantee the success of mobile agriculture applications. 1.2 Motivation This dissertation was proposed by Fraunhofer Portugal Research Center for Assistive Information and Communication Solutions (Fraunhofer AICOS) and performed there. The project consists on designing a mobile solution adapted to farmers’ needs in rural areas of Portugal. In Portugal there’s a lack of research literature regarding farmer’s needs and expectations when interacting with mobile agriculture applications, thus it’s relevant to conduct a research, using a user-centred design (UCD) approach within the field of human–computer interaction (HCI) aiming to develop and adapt the ICT system based on the users’ needs as well as the context of use. 1 Introduction 1.3 Goals The ultimate goal is to better understand farmers’ needs and technology acceptances through user research and analysis and it’ll also be necessary to: •Conduct thorough research on visual aspects to inform the design of graphical user interfaces (GUIs); •Develop prototypes that will explore new and efficient methods of visualization and interaction on Android devices; •Evaluate concrete designs with real farmers; 1.4 Report Structure Besides the introduction, this report has four more chapters. Both chapter 2 and chapter 3 are part of the literature review. In chapter 2 it’s possible to take a look into the use of ICTs in Agriculture. In chapter 3, the process of user-centred design is presented. The chapter 4 presents the followed methodology, divided into phases. In order to conduct this study, the UCD methodology will be divided in four major phases: •Phase One - Requirement gathering: encompasses market analysis of existing solutions, defining the research sample, research methods used and findings. •Phase Two - Analysis: entails scenarios and personas creation, specification of prototype’s requirements with different levels of abstraction. •Phase Three - Design: where according to the specified requirements different guidelines and design principles are taken into account in order to support design decisions in the following next phase. •Phase Four - Prototyping and Evaluation: refers to the creation of low and high fidelity prototypes. In this phase usability tests are performed by real users. Lastly, in chapter 6, the conclusions of this study are presented. 2 Chapter 2 Information and communications technologies in Agriculture With the increased use of Information and Communications Technologies (ICTs) in the agriculture field various areas of opportunity arise. This chapter is composed by two main topics: mAgriculture and Precision Agriculture. Section 2.1 will address the various types of mobile solutions for agriculture, examples of those in the market, and the potential of mAgriculture. Section 2.2 will address the history of Precision Agriculture, technologies used, the benefits, concerns and drivers of adoption of this management approach. 2.1 mAgriculture E-Agriculture is an emerging field focused on the enhancement of agricultural and rural development through the use of ICTs. The mobile phone has become one of the most used ICT devices globally and with the increase of it is penetration in developing countries’ markets, as presented in Fig. 2.1, there’s an opportunity to impact and empower small holder farmers in rural areas. Therefore mAgriculture or mAgri is a subset of E-Agriculture and is the research area concerned with the improvement of the economic environment through the use of mobile computing. Since the introduction of tablets and the higher adoption of smartphones and tablets by farmers, mAgriculture has also become relevant in developed countries [WEU12]. With the use of mobile applications it is possible for small holder farmers in rural or underdeveloped areas to have access to a number of services, from simple tasks like checking the weather [Acc17,Agr16,Bar16] to more complex ones like pests and disease diagnosis [Esp16,PEA17]. It is important to understand the context in which mAgri has evolved and adapted to user’s needs. Some mAgri solutions rely on Short Messaging Services (SMS) 2or Unstructured Supplementary Service Data (USSD) protocols [air15], sending brief messages to communicate with 1http://data.worldbank.org/indicator/IT.CEL.SETS.P2?end=2015&locations=XL&start= 1960&view=chart 2https://www.mfarm.co.ke/ 3 Information and communications technologies in Agriculture Figure 2.1: Mobile cellular subscriptions (per 100 people) in Least Developed Countries: UN Classification 1 farmers. These require less bandwidth, making them less expensive than a phone call but are limited by other factors such as low literacy and lack of knowledge [WSC+16]. SMS compared to USSD can have a higher cost and be a slower service and also has a restriction on the number of characters [Ira10]. Other solutions rely on Interactive Voice Response (IVR) 3. This can be a great way to overcome the obstacle of illiterate users. All these technologies can be combined and some proposed solutions encompass SMS, USSD and even IVR all in the same service, while other services evolve and adapt to their costumers’ needs, like the Airtel Kilimo service that evolved from a SMS & IVR solution to USSD only [air15]. Even though mAgri services are supposed to empower smallholder farmers it’s hard to do so if farmers aren’t aware those services or if they do not know how to interact with them. A research study conducted in rural Kenya [WSC+16] indicated that novice or low-literate users preferred making voice calls rather than sending text messages due to their limited technological proficiency. Other variations of mAgri services are the use of call centres and video calls to better instruct farmers. Even more advanced, and something developed countries are investing more on, is the mobile web 4and installable applications [Ava13], these allow infinite possibilities of integration with other devices and technologies and provide a much more sophisticated interaction. Findings from another research [GWO+16] conducted in Kenya involving 56 farmers evaluating an mAgri application that contained SMS, Voice and web-based components indicated that all the interviewed farmers mentioned that the delivery platform (i.e., IVR, USSD, SMS) would determine if they would use the system, being SMS the preferred mode of interaction. Therefore several external factors should be taken into account when developing ICT solutions for farmers, for instance access to electricity, network coverage, users education level, cultural, social and gender constrains [M+16]. The choice of technology should be adapted to the users needs, kind 3http://farmerline.co/ 4https://www.plantvillage.org/ 4 Information and communications technologies in Agriculture of devices they have access to (see figure 2.2 relating mobile phone type to technologies) and capabilities [GWO+16]. Figure 2.2: Personal mobile devices and delivery technologies [HGWH13] 2.1.1 mAgriculture Solutions The GSMA mAgri Programme 5was founded in 2009 and works with the mobile operators, who are members of the GSMA, together with agricultural organizations and development organizations to look at where mobile technology can bring benefits to the agriculture sector. GSMA proposes [Int15b] that mAgri solutions should be divided into three categories: 1. Agricultural value added services (Agri VAS) — These consist of machine-to-human interaction and are delivered through SMS, USSR, helplines, interactive voice responses (IVR) and, increasingly, by web and mobile applications. Even though farming is one of the main drivers of the economy in developing countries there’s a key challenge that arises, the productivity level of farmers in developing countries is lacking when compared to that of developed countries. 2. Machine-to-machine (M2M) — This category is, as the name indicates, related to technology that coordinates multiple machines, devices and appliances connected via communication channels. This category features devices like smartphones or tablets and Precision Agriculture is an example of an application of M2M in agriculture (this concept is covered in detail in section 2.2) [Int15a]. 3. Mobile financial services tailored for the agricultural sector — This third category consists of solutions in Agri MFS (Mobile financial services), combating farmer’s financial 5http://www.gsma.com/mobilefordevelopment/programmes/magri 5 User-centred Design subsequent adoption and making sure that even after deployment the product will continue to be improved. Gould and Lewis [GL85] state that three principles should be followed in order to lead to a “useful and easy to use computer system”: •“Early focus on users and tasks” — studying and observing them and involving the users since the beginning of the design process; •“Empirical measurement” — recording and analysing users’ performance and reaction; •“Iterative design” — when problems are encountered, fix them and continue testing. These principles are timeless and now accepted as the basis for a user-centred approach [PSR15]. 3.1 Disciplines surrounding UCD 3.1.1 Human-computer interaction Human-computer interaction (HCI) is an interdisciplinary field that is defined as “a discipline concerned with the design, evaluation and implementation of interactive computing systems for human use and with the study of major phenomena surrounding them.” [DFAB03]. It has become an umbrella term that encompasses diverse disciplines. HCI is a subset of usability (see 3.2) and focuses on how humans interact with computing products while UCD is a methodology used to ensure that the products meet the users’ needs [Low13]. 3.1.2 Ergonomics The International Ergonomics Association establishes that “Ergonomics is the scientific discipline concerned with the understanding of interactions among humans and other elements of a system, and the profession that applies theory, principles, data and methods to design in order to optimize human well-being and overall system performance.” 1 Ergonomics is traditionally a study of the physical characteristics of interaction [DFAB03], is a field closely related to HCI but distinct, and helps constraining the design of systems based on user performance and the role of the interface in that performance. It goes from how controls are designed to the physical environment in which the interaction takes place. 3.1.3 Interaction Design Interaction Design (IxD) is about “designing interactive products to support people in their everyday and working lives” [PSR15]. Winograd [DM97] described it as “the design of spaces for 1http://www.iea.cc/whats/ 12 User-centred Design human communication and interaction”. IxD encompasses academic disciplines like engineering, informatics, amongst others and also design practices such as graphic design and product design. HCI is an area of study opposed to IxD, a design craft that deals with interaction, not necessarily being specific to computers. The process of interaction design goes through the identification of user needs, developing alternative designs to fit these, building interactive prototypes that can be communicated and assessed, and evaluating throughout this process what is being built and the user experience it offers. 3.2 Usability and User Experience Usability Nowadays it’s common to hear the term “user-friendly” to connote ease of use and interaction, but “in reality, different users have different needs, and a system that is “friendly” to one may feel very tedious to another” [Nie94]. Usability is considered by Jakob Nielsen as a multi-dimensional property that refers to how easy interfaces are to use 2defined by five quality components [Nie94]: •Learnability — a system should be easy to perform basic tasks by a first time user; •Efficiency — once the user has learned the system he should be able to be highly productive; •Memorability — after a period of inactivity a user should be able to pick up easily and re-establish proficiency; •Errors — the system should have a low error-rate and allow users to easily recover if they make errors; •Satisfaction — the system should be pleasant to use. “if the system forces the user to adopt an unacceptable mode of work then it is not usable” [DFAB03] Users should not only find a system efficient and effective, they should also want to use it, and that comes with the development of interactive products that ensure usability, and most importantly, that measure it. User Experience User Experience (UX) is defined by Don Norman and Jakob Nielsen through the sentences: “The first requirement for an exemplary user experience is to meet the exact needs of the customer, without fuss or bother. Next comes simplicity and elegance that produce products that are a joy to own, a joy to use. True user experience goes far beyond giving customers what they say they 2https://www.nngroup.com/articles/usability-101-introduction-to-usability/ 13 User-centred Design want, or providing check-list features. In order to achieve high-quality user experience in a company’s offerings there must be a seamless merging of the services of multiple disciplines, including engineering, marketing, graphical and industrial design, and interface design” [NN]. UX and usability should be distinguished, the latter term being a quality attribute of the UI, related to how easy a system is to use, how efficient and so forth. UX is an even broader term, related to how the person feels when using a product. Some examples of those feelings can be excitement, fun or aesthetic pleasure. 3.3 Understanding the user 3.3.1 Problem Space In order to solve any problem it’s necessary to understand the problem space, therefore ask what, why and how [PSR15]. Identifying usability and user experience goals are a part of that process, being those of an existing product, how it can be changed or improved, or those of a new one. Having a good understanding of the problem space can help inform the design space what kind of interface, behaviour and functionality to provide. 3.3.2 Conceptual Models When designing a system is crucial to, in the early stages, understand and conceptualize the user experience. It’s necessary to have a good understanding of the problem space, meaning what you want to create, why and how it will support users in the intended way. A conceptual model is “a high-level description of how a system is organized and operates.” [JH02]. It helps to outline what people can do with a product and what concepts are needed to understand in order to interact with it. Thus it includes all concepts exposed to the user and the sequence of operations performed to complete a task. The better the conceptual model matches the user’s existing mental models the better the user experience will be. 3.3.2.1 Metaphors Metaphors can be used to convey the understanding of how an interface works and help users interaction. They provide a way to teach new concepts and familiarise the user [DFAB03], helping them understand the underlying conceptual model [PSR15]. They should be used with this purpose but sometimes can have negative effects, being important to understand when the use of a metaphor breaks design principles. 14 User-centred Design 3.3.2.2 Interaction types There are different ways a person interacts with a product and Jenny Preece, Helen Sharp and Yvonne Rogers [PSR15] propose there are four different types of interaction that are not meant to be mutually exclusive nor meant to be definitive. •Instructing — issuing commands using keyboard and function keys and selecting options via menus; •Conversing — interacting with the system as if having a conversation; •Direct Manipulation — interacting with objects in a virtual or physical space by manipulating them; •Exploring — moving through a virtual environment or a physical space. 3.3.3 Mental Models Mental models can be described as internal constructions of reality made by the mind that it uses to anticipate events, describing the user’s view of how things work [Cra43]. The difference between conceptual models and mental models is that conceptual models are “devised as tools for the understanding or teaching of physical systems” [RB88] and mental models are “what people really have in their heads and what guides their use of things” [RB88]. 3.3.4 Cognition Interacting with technology involves a lot of cognitive processes. Cognition can be distinguished between two general modes [Nor93]: Experiential Cognition, a state of mind in which we perceive, act and react to events around us effectively and effortlessly (e.g. driving a car), and Reflective Cognition that involves thinking, comparing, and decision-making (e.g. learning or writing). 3.3.4.1 Cognition Processes Some of the processes that affect behaviour are: •Attention — Information structure at the interface that better captures user’s attention, e.g. use of perceptual boundaries, colour and sound. •Perception and recognition — Perception refers to how information is acquired. Obvious implication is to design representations that are readily perceivable, e.g. text should be legible and icons should be easy to distinguish and read. •Memory — Humans recognize things better than recalling, for instance we’re better at remembering images than words. Interfaces should help people remember essential information, promoting recognition rather than recall. Memory involves encoding and then 15 User-centred Design retrieving of knowledge. Context is important when it comes to memorization, sometimes it can be difficult to recall information encoded in a different context. •Reading — Reading can be quicker than speaking or listening but some users avoid reading [CR87]. It’s easier to recognize word shapes and because lower-case words have more distinct shapes, they are quicker to read. 3.4 The Process of Design According to Alan Dix et al. [DFAB03] there are four main phases plus an iteration loop relating to the design of interaction, represented in Figure 3.1. The first being the “What is wanted” - requirements phase (3.4.1), then the “Analysis” phase (3.4.2), after that the “Design” phase (3.4.3), a loop of “Prototype” - evaluation, iteration and prototyping phase (3.4.4) and finally the “Implement and deploy” phase (3.4.5). Figure 3.1: Interaction design process [DFAB03] Life cyle models It’s important to have an understanding about how the design process can fit in the software development process. Life cycle models help show how activities are related to each other, as they are management tools and simplified versions of reality. Traditional models are more adequate for smaller teams, providing a simple process, in contrast to larger systems that need a more detailed and sophisticated model. Here are a few of these models: •Waterfall Model — In software engineering this model was the basis of many development cycle in use nowadays. This traditional view consists of a number of processes in a sequential fashion. •Spiral Model — The Spiral Model [Boe88] takes into account prototyping throughout the design process. The spiral process follows the main stages described in the Waterfall Model, but does it iterating several times, checking and evaluating ideas and enabling risk analysis. 16 User-centred Design •The Star Model — The Star Model [HH89] also features an iterative nature, but with Evaluation at the centre stage of activities. There’s also a lack of ordering, making it possible to go from each activity to the other, provided you go through Evaluation in the process. This model is a reflection of empirical studies conducted on how designers went about their work. 3.4.1 Requirements Requirements are obtained through data gathering and data collection with the help of certain techniques and activities to get a better understanding of users’ needs. Conventionally, requirements can be divided into two types, functional and non-functional [Ben10]. According to Benyon functional requirements relate to what the system must do while non-functional requirements are a quality that the system must have and these include for instance usability, aesthetics and cultural acceptability. 3.4.1.1 Research Methods In order to capture requirements it is important to understand users and conduct user research. To define the research goals, a look should be taken at the existing issues that should be presented as goals. After, these goals should be prioritized and finally rewritten as questions to be answered [GKM12]. There are several methods of design research, and according to Dan Saffer [Saf09] these can be roughly divided into three categories: observations, interviews and activities, where the later one includes subjects performing and self-reporting their activities. Observations Observing what people do can be one of the most elucidating methods of design research. Subjects can be observed without any interference from the designer or with interaction, like asking them questions about what they are doing and most importantly, why. Some ways of doing observations include: •Fly on the wall — unobtrusively observing users’ behaviour on location [Saf09]. •Shadowing — following subjects while they go about their daily activities [Saf09]. •Undercover agent — by posing as a normal person in an environment it’s possible to observe people interacting as they naturally would [Saf09]. •Ethnography — a form of observation deriving from anthropology [DFAB03]. Ethnographers observe users in their natural environment, making it possible to analyse the behaviour of individuals in the context of use. This process can take hours to months and the goal of 17 User-centred Design this observation is to collect data that’ll influence the interface design process [SP03]. Compared to other approaches to data gathering Ethnography has the aim of observing without “imposing any a priori structure or framework upon it” [PSR15]. •Contextual inquiry — also studies the user in context but it differs from pure ethnographic study due to the intention: “the intention is to understand and to interpret the data gathered, and rather than attempting to take an open-ended view, the investigator acknowledges and challenges her particular focus” [DFAB03]. This is a variation on shadowing because the subjects are asked questions [Saf09]; •Participatory Design — in this approach the user comes out of their work situation, either mentally or physically and becomes apart of the design team, entering into the subjective experience of the workplace [DFAB03]. Interviews Talking to users and hearing their stories can provide a great insight into their experiences and attitudes but can also be dangerous when performed alone because what people say and do are typically very different [Saf09]. •Directed storytelling — asking subjects to talk about how they performed an action or their experience in a specific time [Saf09]. •Focus group — by assembling a group of representative users in the same group can be helpful to gain a consensus view and to get a better understanding about areas of conflict. •Unfocus group — a variation of Focus groups, developed by IDEO 3where extreme people passionate about the product/service are brought together allowing to explore the subject from different viewpoints [Saf09]. •Role playing — consists of acting out different scenarios by a group or individual in order to perceive how a subject/product/system affects users’ emotions and attitudes [Saf09]. •Extreme-user interviews — another method created by IDEO 4requires interviewing people on either extreme of the spectrum, not only the big broad mainstream [Saf09]. This method is used to spark creativity and spur new thinking. Activities In design research, doing activities allows designers to go beyond observing and talking to subjects. There are a number of activities proposed by different authors and each can be a great tool 3http://www.ideo.org/ 4http://www.designkit.org/methods/45 18 User-centred Design depending on the project and design goals. Collaging 5and drawing 6are creative activities, allowing subjects to be more involved by making and explaining their creations and allowing designers to get a better understanding of their values and thought process. Card sorting 7can also be a quick and easy method of understanding, by having concepts written or printed on cards and the participant arranging, grouping, categorising or organizing them. Self Reporting With this approach subjects, without the direct involvement off designers, record their activities and thoughts, allowing them to feel more comfortable and less embarrassed, and then the researcher collects and analyses the findings •Journals — the subject journals particular activities [Saf09]. •Beeper studies — the subject is given a beeper and when the designer sets it off during the day the subject records on a journal what they were doing at the time [Saf09]. •Photo/video journals — the subject is given a camera and is told to document his life or for instance to take a picture of people in his trust-circle or who influence him 8. Other data gathering techniques for establishing requirements are the use of questionnaires, studying documentation and researching similar products [PSR15]. Questionnaires can be made up off open-ended or close-ended questions and therefore can be a more qualitative or quantitative technique, depending on what is pretended. Studying documentation is a technique that doesn’t require the involvement of stakeholders and can be a good source of data about steps and regulations governing a task. Lastly, researching similar products can be helpful to get a better insight on features and interactions already provided in the market and to prompt requirements. 3.4.2 Analysis The results from the methods and activities previously referenced need to be analysed in order to bring out key issues that’ll guide the Design Phase. 3.4.2.1 Scenarios According to Alan Dix et al. [DFAB03] Scenarios are stories or narratives of interaction and they’re perhaps the simplest design representation. They can be very powerful and can go from a simple description to a full on script describing the situation or context of interaction. Scenarios can also be augmented by sketches, storyboards or other mock-ups and force the developing team to act out potential patterns of use and to anticipate possible problems that could be encountered 5http://www.designkit.org/methods/25 6http://www.designkit.org/methods/49 7http://www.designkit.org/methods/24 8http://www.designkit.org/methods/65 19 User-centred Design later on. There are other applications for scenarios, for instance telling stories in a natural way allows stakeholders to get a better understanding of users, with the use of non-technical vocabulary and phrasing, and be more participative in the development process [PSR15]. “The basic argument behind scenario-based methods is that descriptions of people using technology are essential in discussing and analysing how the technology is (or could be) used to reshape their activities. A secondary advantage is that scenario descriptions can be created before a system is built and its impacts felt.” [RC01] Rosson and Carroll (2002) [RC01] explain an approach that illustrates the use of scenarios within a usability engineering framework. Distinct types of scenarios are illustrated in Fig. 3.2, showing: •Problem Scenarios — these tell the story about how users currently engage in their activities and enumerate what features of the current situation have an important consequence. They do not necessarily “emphasize problematic aspects of current practices, but rather [...] describe activities in the problem domain.” 9; •Activity Scenarios — after the staring point provided by problem scenarios, activity scenarios are related to how the user could possibly be doing their activities, “maintaining or even enhancing the positive consequences for the actors while minimizing or removing the negatives” [RC12]. •Information and Interaction Design Scenarios — “After completing the claims analysis, the goal of information design is to specify representations of a task’s objects and actions that will help users perceive, interpret, and make sense of what is happening. The goal of interaction design is to specify the mechanisms for accessing and manipulating the task information and activities.” 10 3.4.2.2 Personas Related to scenarios it is also possible to define personas. A persona is an invented person used to represent a type of user inserted in a scenario. Assigning to each persona a name, face, personality and possibly other information, allows to talk concretely about a potential end-user. These can be very helpful. The difference between Scenarios and Personas is that the first one describes the use of a product or an activity to achieve a goal, while the latter characterizes a typical user, defining who the story is about, their main motivations, goals and pain points (see Figure 3.3). 3.4.2.3 Task Analysis Task Analysis can contribute to the statement of requirements when creating a new system or can help in the production of training materials and documentation when existing systems are analysed. Alan Dix et al. [DFAB03] describe three different approaches to task analysis: 9http://ldt.stanford.edu/~gimiller/Scenario-Based/ProblemScen.htm 10http://ldt.stanford.edu/~gimiller/Scenario-Based/Infor-InteractDesign.htm 20 User-centred Design Figure 3.2: Overview of scenario-based design (SBD) framework [RC12] Figure 3.3: Relationship between Scenario and Persona 11 •Task decomposition — “which looks at the way a task is split into subtasks, and the order in which these are performed”. Hierarchical task analysis (HTA) is an technique that provides an understanding of the tasks users need to perform to achieve a certain goal and can be recorded in a tree diagram or a textual outline format (see Figure 3.4). Figure 3.4: Hierarchical task analysis example for ordering a book 12 11https://www.smashingmagazine.com/2014/08/a-closer-look-at-personas-part-1/ 12http://www.uxmatters.com/mt/archives/2010/02/hierarchical-task-analysis.php 21 User-centred Design 28 Chapter 4 Methodology The purpose of this research is to come up with a mobile solution, while having farmers at the centre of the design cycle. The previous chapters introduced literacy relating to ICT in Agriculture and UCD, both crucial to guide the following research methodology. As such, the following phases presented are based on the process of design presented in 3.4. As previously mentioned in chapter 2there are various types of mAgri solutions, from relying on SMS or USSD protocols to more complex ones, such as web and installable applications. In this study, the choice of technology was requested by Fraunhofer AICOS. It should be an android installable app, because of the interest in researching a solution with a higher level of interaction. 4.1 Phase One - Requirements gathering The first phase, as described in the previous chapter (see 3.4.1), encompasses various methods and activities to gather requirements. 4.1.1 Market Analysis Different users and stakeholders are crucial to this phase, but due to a delay in getting in contact with important organizations and associations, that would be the bridge between the research team and farmers working in the field. The first step of this phase was researching what the market already offered in regards to Agriculture Apps. As mentioned in section 2.1.2, a study developed by Gary Woodill and Chad Udell dating back to 2012 pooled and categorized mobile agriculture applications existing in the market to that date. This study is considered outdated nowadays and also was restrictive, only including solutions created towards the Canada and United States markets. To get a better understanding of what is offered nowadays to the general public a new updated study was made on agriculture apps in different categories, such as, Learning and Reference Apps, Farm Management Apps, Diseases and Pests Apps and others (presented in Appendix A), gathering applications available up until March of 2017 on the Google Play Store. The sample analysed in this updated study includes 33 mobile android applications and that 29 Methodology sample was selected by: analysing the most popular apps in different categories of mAgri. The categories selected are the ones referenced in the outdated study, minus the category “Conference Apps”, because it was not relevant to this study. The sampling was made to include apps from different locations, supporting different languages and ranging from multiple features, in different categories, to single purpose apps. This allowed for a diverse sample and for a broader view of what is available in the market nowadays. After collecting various data about each app, such as: number of installs, rating, age, supported languages, source location, and some others, it was also important to assign to each app the categories that best fit the functionalities marketed or provided by the developer. Due to the difficulty found assigning this categories, because the apps usually provided multiple functionalities, and sometimes cross-cutting ones, various professionals with distinct backgrounds in areas such as Design, HCI and Software Engineer participated in an activity where they would assign different features to “super-categories” (see Fig. 4.1). A more in-depth report on the activity is provided in Appendix B. Figure 4.1: Categorizing activity Here are some relevant findings from the study: Figure 4.2: Findings of Market Analysis The most popular categories are Weather Apps, Farm Management Apps, Market Data Apps and Learning and Reference Apps. By far the most common location is India followed by the 30 Methodology USA. Regarding languages, English is predominant, being present in 29 out of 33 apps and more apps are monolingual than multilingual. 4.1.2 Research sample This research was performed in Portugal and because of the need to involve the potential end-users in the design process the research sample consists of farmers in rural areas of Portugal. A study performed in 2013 by the Statistics Portugal institution [INE13,INE15] indicates that agricultural producers in Portugal are the oldest of the the European Union (EU), with more than 52% of them being aged 65 years old or more. The vast majority (70%) only has basic qualifications and only practical agricultural experience. Little of them live exclusively from farming, supplementing their income with retirement pensions. Another important factor to take into consideration when looking at mAgri solutions for different areas of the world, specially in developing countries or rural areas, is the adoption rate of smartphones in that location. The Google Consumer Barometer Report 1points to an increase in the adoption of smartphones in Portugal over the past 5 years, with a 59% rate in 2016 as illustrated in Fig. 4.3, compared for instance to 33% in India in 2015 and 74% in the United Kingdom in 2016. With a sample size n=1,000 in each country except India (n=4,000) the target population consisted of nationally representative total population (online & offline) aged 16 years old or more. Survey data was collected by TNS Infratest via telephone with a dual-frame approach (landlines and mobile phone numbers) 2. Another study conducted by Marktest indicates that the penetration rate of smartphones in Portugal has grown over 89% since April of 2013. Smartphones exceeded feature phones back in October of 2014, counting as of February of 2016 with 68% of the total number of phone users, as illustrated in Fig. 4.3. This study was composed of a sample of individuals aged 10 or more, residing in Portugal and in the Autonomous Regions of Madeira and Azores, collecting 1250 interviews monthly, 1000 from homes with a landline telephone and 250 from households with a mobile service, without wireline voice traffic. Agriculture encompasses various types of farming and in order to collect requirements from real farmers it was necessary to focus on a specific area of agriculture instead of researching within the large scope of Agriculture. Since this study was conducted focusing in rural areas of Portugal it was relevant to focus on an agricultural activity that would be prevalent in these areas. When it comes to the size of agricultural holdings, according to an agricultural census “with the exception of farms with farm type "Various crops and livestock combined" (14.7%), holdings specialised in vineyards are the most common in Portugal” 3. Therefore the scope of the study went from Agriculture in general to Viticulture in the Douro region, which has a low degree of urbanization, as illustrated in Fig. 4.4. 1https://www.consumerbarometer.com/en/ 2https://www.consumerbarometer.com/en/about/ 3http://ec.europa.eu/eurostat 31 Methodology (a) Google Consumer Barometer - Percentage of people who use a smartphone in Portugal (b) Marktest study of Smartphone adoption in Portugal Figure 4.3: Smartphone adoption in Portugal Figure 4.4: Degree of Urbanization - Douro Region 4 The Douro wine region was demarcated because of the need to regulate the sector and sales of Port Wine 5, originating the Demarcated Douro Region (DDR), that is subdivided in three subregions, Lower Corgo, Upper Corgo and Upper Douro. According to the latest data provided by IVDP (Instituto dos Vinhos do Douro e Porto) these three regions have a total area of approximately 250.000 hectares, with 43.655 hectares dedicated to vineyards 6. Approxitamely 21.432 4http://ec.europa.eu/eurostat/statistical-atlas/gis/viewer/?config=RYB-2016. json&mids=2,8,9,12,2013-3-NAMES,CNTR-OVL&o=1,0.7,1,1,1,0.7¢er=38.10049,-3. 19915,6&ch=10,11,34,51,58,70,89& 5https://www.ivdp.pt/en/docs/douro_en.pdf 6https://www.ivdp.pt/estatisticas_novo2.php 32 Methodology winegrowers, each owning an average of 2 hectare of land, work in these regions. Therefore small winegrowers play a big role in the production of wine in the Douro region. This has been changing over the years, with the area per producer increasing since 2010, as illustrated in the graph in Fig. 4.5. Figure 4.5: Area per Producer More statistical data about Portugal and the Douro region can be found in Appendix C. 4.1.3 Research Methods As mentioned above, Viticulture was the chosen area of focus for this study, and it deals with the production of grapes with the cultivation of vineyards, and usually, especially in this area of the country, follows more traditional methods due to the impossibility of mechanisation, because of the difficult accessibility present in the area or due to a need to keep the heritage of the process alive. The methods used to capture requirements were chosen from the ones mentioned in the previous chapter (see 3.4.1.1). This choice depended on time constrains, availability from the people involved and also on which methods would provide a better insight on the viticulture activity, because none of the research team members was an expert on this field. 4.1.3.1 Interviews Contact was made with three different entities directly involved in the wine growing business, each with different roles and goals. Interviews were conducted with winegrowers and relevant stakeholders, such as winemakers, lab and field technicians, cooperative and association members, thus providing insight into the main problems encountered, by different people, with different roles. On total seven interviews were conducted, some on-site, while others went via skype or phone call. These interviews helped to identify key-areas of opportunity while also helping to educate the research team on the field of viticulture. 33 Methodology One interview consisted of a Focus group activity, with three winegrowers, one technician and two cooperative representatives as participants. This activity was important to cross opinions and find problem areas from different points of view. The interviews began as more unstructured interviews and with the advances of the project they transformed into more semi-structured and structured, being composed of both open and close ended questions, due to the increase in specificity in relation to the problems encountered and the matters addressed. An example script can be found in Appendix D. 4.1.3.2 Observations Field visits are fundamental to better understand the needs of the winegrowers and associated stakeholders, the problems they encounter, how they solve them and how technology may facilitate that problem solving. Two field visits were conducted, one to a smaller winegrower and one to a medium/large winegrower, making it possible to see different realities of the same type of agriculture. It was clear to see, during both visits, that the access to the plots is tricky for winegrowers, especially for an older person, and for machinery. The terrains have to be adapted for tractors and other machines to access it, as can be seen in Fig. 4.6. The conditions under which the farmer’s daily tasks are done can then be very difficult, especially during warmer days, due to the dry and warm climate of the region. Figure 4.6: Vineyards on steep slopes and terraces It is important to highlight that the visits were conducted at the beginning of the year, in April, making it hard to get an understanding of the whole winegrowing cycle. During this time of year winegrowers are preparing for the flowering phase and being less active when compared to the months anticipating the harvest time. 34 Methodology Figure 4.7: Winegrowing Calendar 7 Due to the difficulty of having an insight into all the activities winegrowing requires, methods that only allowed observing users were not viable, therefore the Contextual Inquiry technique allowed to, not only observe users on their natural setting, but also to conduct an interview in the field and guide the conversation topic to relevant topics that were previously mentioned in interviews. 4.1.4 Findings Collecting the findings from interviews and visits resulted in three main themes that could be considered areas of opportunity: •Forecasting the harvesting potential — predicting the quantity of grape/harvesting potential is useful for a better management of resources, aiding in better decision making of when to harvest and how long it will take. Nowadays one of the methods for production forecast, performed in May, is the Pollen Method. “This model involves the capture and analysis of the amount of pollen released by grapevines in three locations representative of the three subregions of the Demarcated Douro Region (RDD), thus integrating climatic and phenological data.” 8This method is sometimes too precocious and can result, at a later stage, in very different outcomes. •Ripening control — besides quantity, the quality of the grapes is crucial to achieve a high quality wine. This also influences when the harvest takes place and this analysis is made by collecting grape berries and testing in a lab a list of parameters, such as: ph level, acidity, 7http://www.evineyardapp.com/blog/wp-content/uploads/2016/02/vineyard-calendar. png 8http://www.advid.pt/imagens/boletins/13753484584274.pdf 35 Methodology among others. The process of collecting samples, labelling them, taking them to the lab, performing tests and reporting the results can sometimes not be very organized, be laborious and time consuming. •Pests and Diseases Prevention and Control — This is an issue that affects both the quantity and quality of grapes. Pests and Diseases may generate very large losses, not only for that year’s harvest but also for future harvests. An overview of the main pests and diseases that attack the vineyard in the Douro region is provided in Appendix E. Prevention and Control are conducted depending on the pest or disease in question. Usually risk assessment is performed, for pests, through the use of traps (traps examples illustrated in Fig. 4.8) and for diseases depending on the knowledge about its biology and epidemiology. Control is conducted through cultural precautions and/or chemical control. Figure 4.8: Two types of traps: Yellow sticky trap and Pheromone trap From these three opportunities “Forecasting the harvest potential” was considered the less valued to develop an ICT solution for, because the level of interaction between the farmer and the application would be low considering the methodology followed nowadays, therefore two themes remained. In order to chose which opportunity of design would be further investigated and the area our solution would tackle, a brainstorming activity was conducted, with the same participants from the activity mentioned in 4.1.1. During this activity (see Fig. 4.9) an introduction and presentation of the context of the project were presented and in order to provide a better understanding of data gathered to that date, personas and scenarios were also presented (see Appendix G). The purpose of the brainstorming activity was also to generate concepts and ideas for a mobile app and resulted in possible features for these two areas of opportunity. It was established that the main area of opportunity was “Pests and Diseases Prevention and Control” because it would allow to better study interaction of the winegrower, contrary to “Ripening control” that would result in an application mostly used in a more controlled environment, inside a laboratory, and by people that could be easily guided on how to interact with the application. A more detailed report on the activity can be found in Appendix F. 36 Methodology Figure 4.9: Brainstorming activity Concerns Some of the main concerns relating to “Pests and Diseases Prevention and Control” are: •Prevention and treatment can be instinctive — usually less educated users go for what they know and their experience tells them, making sometimes decisions that can be fatal to their crops. One example found during an interview was of how winegrowers sometimes choose pesticides because the name of the product is long and difficult to pronounce and not because it was more effective. •Late treatment can be devastating — timing is crucial when trying to attack pests and treat diseases. Sometimes winegrowers perform treatment when it is more favourable to them. Some perform treatment on the weekend because they can get family to help and save money on hiring people. Another factor that influences the timing of the treatment is the weather, if it is too hot people treat during the late afternoon and in the early morning. If it is raining heavily treatment has to be postponed. •Selective prevention/treatment can be difficult — inside the same plot many factors influence how a disease progresses and how pests attack, some factors being: altitude, sun exposure, proximity to a body of water, the planted grape variety, among others. Therefore it is important for treatment and evolution tracking of the vineyard to be done selectively, meaning, treating spots instead of the whole plot. Goals Summarizing, when it comes to pests and diseases prevention and control it is important to get all the information needed to know how, when and where to act, in order to prevent losses. Besides this, some other goals were mentioned by different stakeholders: •Increase farmers’ productivity — by reducing the amount of paper work and by computerising, organizing and structuring information in one place. Also, as pointed out before (see 4.1.3.2) because the access and work on the field can be exhausting (due to terrain or 37 Methodology Figure 4.15: Use Cases Diagram 44 Methodology Figure 4.16: Hierarchical task analysis 45 Methodology 4.2.4 Non-functional requirements The representation models mentioned above refer to Functional requirements, a concept that is explained in 3.4.1. Listed bellow is a set of non-functional requirements that also impact the user experience: •Availability: the application must be available at all times (both offline and online); •Documentation: a quick-reference guide should be provided; •Efficiency: the application must have the lowest possible response times whether we are talking about the Interface response time (must feel as native as possible) or the response times in the communication with the external services. Because of the context of use, performance should be adjusted to conserve battery power; •Extensibility: should be possible to extend the system to support more pests and diseases relating to vineyards; •Interface: the user interface must be very intuitive so that the user can access all the functionalities with minimal interaction with the software. The interface must also be responsive to all kinds of mobile devices’ screen sizes. Screen luminosity should adjust to the environment.; •Maintenance: the application should be built in such a way that it is easy to modify the implemented functionalities. 4.3 Phase Three - Design Once the requirements have been established, the Design Phase begins. In this phase it was necessary to collect the rules, guidelines and design principles that would inform the creation of the concrete designs, that being, the prototypes. This selection was made accordingly to the target of users and the context of use. Even though the requirements in section 3.4.1 present two different actors, the winegrower and the expert, the design guidelines and prototypes developed relate only to the farmer. This decision was made due to the focus of this study being on farmers’ interaction with a mobile application. 4.3.1 Material Design In order to provide a “unified experience across platforms and device sizes” 10 Google 11 created an innovative design language based on classic design principles. Material’s principles and guidelines served as the foundation for the design of this mobile solution, adding to these, guidelines from other authors. 10https://material.io/guidelines/#introduction-goals 11https://www.google.com/intl/en/about/ 46 Methodology Material lists three user interface design principles 12 to their philosophy: •“Material is the metaphor” — by playing on the behaviour of paper and light, material allows the user to have a perception of objects and surfaces that is natural and intuitive. This metaphor provides an inherent understanding about the separation of things and relationship of things on the interface. •“Bold, graphic, intentional” — means having bold typography and color choices, edgeto-edge imagery and intentional white space to create an immerse experience for the user. •“Motion provides meaning” — motion provides increase of visibility and feedback. When it comes to guidelines, Material provides an extensive documentation 13 for several aspects of the design, such as, style, layout, components, gestures and others. These guidelines depend directly on the solution’s requirements, therefore the extracted concepts from Material are described in detail in section 4.4. 4.3.2 HCI4D Guidelines The ICT4D Competence Center (ICT4DCC) at Fraunhofer AICOS, that aims to investigate, research and develop mobile solutions for areas such as: Agriculture, Health, and Citizens & Government collected a set of research based guidelines to tackle illiteracy or low literacy, either textual or technological, in developing countries [DMG+,Gie15]. Even though Portugal is considered a developed country, some of those principles are transversal and can be helpful to inform design decisions in this study. As mentioned in Appendix C, dating back to 2011 the Douro region had an illiteracy rate of 3.74%, with a deviation of 0.4% from the whole country. This can be an indication that for this target users, guidelines for semi-literate users can be more pertinent than those for illiterate users. A list of relevant HCI guidelines for this research are presented next. 4.3.2.1 Ergonomic Factors 1. “The avoidance of multi-function buttons is optimal for users with little experience” — a button should only provide one functionality because multi-function buttons confuse users with little experience. 2. “Over-cluttered buttons should be avoided” — for examples alphanumeric keypads should be avoided because they provide issues regarding access and control. 3. “Double-tapping is problematic to low-literate users and should be refrained” — one click actions are more effective and have the highest recall. 12https://material.io/guidelines#introduction-principles 13https://material.io/ 47 Methodology 4.3.2.2 Interface Design •Text 1. “Illiterates and semi-literates have different requirements” — the advice for fully illiterate users is to use minimal text but it should be kept in mind that this could be prejudicial for semi-literal users. For semi-literal users text has been show to reinforce semi-literates reading skills. 2. “Augment text with other modalities” — using audio and images can provide illiterate users a replacement for text and increase ease of interaction for semi-literate users. •Graphics 1. “Use hand-drawn graphics instead simplified abstract or photorealistic representations” — hand-drawn cartoons are preferred and more easily understood by illiterate or semi-literate users. 2. “Do not use icons on their own” — using textual descriptions next to icons allows for an easier understanding of the functionality provided. 3. “Use icons and graphics understood by the target culture” — icons should reflect the culture and experiences of the target-users. 4. “Indicate that icons can be selected” — icons’ behaviour should change when selected to convey that they’re selectable. 4.3.2.3 Navigation and Information Architecture 1. “Avoid hierarchical structures” — for low literacy users linear navigation structures should be favoured over hierarchical structures because they’re easier to understand. 2. “Encourage exploration of the interface” — by providing feedback, undo functionalities and confirmation dialogues after high risk operations, the user feels less nervous to explore. 3. “Keep the home screen simple” — in order to avoid intimidating novice users the home screen should be kept simple. 4. “Minimize the number of tasks possible on each screen and offer only one way to accomplish them” — offering more than one way to perform a task might confuse some users. 5. “Avoid scrollbars” — some users do not realize that functionalities appear by vertical scrolling. 48 Methodology 4.3.3 Designing for Older People As mentioned in section 4.1.2 most farmers in Portugal belong to an older age group, making it important to look into guidelines for designing for older people. The following guidelines are a selection of relevant recommendations: 1. Colour — elderly people are less sensitive to colour contrast, therefore it is harder for them to distinguish some colours [Phi11]. 2. Buttons — bigger buttons with larger touch areas may give better performances but the size of the button should be considered based on how it’ll affect other elements, such as, the screen layout [Phi11]. For example, the designer may need to split working displays or make use of scroll bars. 3. Text size — when it comes to font sizes, 14-point size or higher is recommended and so is the use of black text on white background. 4.3.4 Context of Use It is also important to take the context of use of the application into consideration. Most of the application’s functionalities need to be performed in the field, therefore outdoor use has some problems associated. In order to use a device under direct sun light contrast is very important [OT11]. Placing controls in a consistent location might make it easier to overcome the problem of fingertips on the screen combined with sun glare 14. If possible, placing the most frequent interactions on the same side of the screen, at a reachable distance, allows the user to perform one-handed interactions. Those reachable areas are illustrated in Fig. 4.17. Figure 4.17: Thumb Zone Mapping 15 14https://www.researchgate.net/post/What_design_factors_should_I_consider_when_ designing_a_mobile_app_for_outdoor_use 15https://www.smashingmagazine.com/2016/09/the-thumb-zone-designing-for-mobile-users/ 49 Methodology 4.3.5 Conflicts Conflicts will naturally arise when taking all these guidelines into consideration, here is a list of some of them: •Gestures — Material relies heavily on touch mechanisms and activities that can provide a richer user experience. A problem arises, as mentioned in item 3, even double tapping can be hard for low literate users. To tackle this issue Material provides a series of instructions on gesture education 16. •Navigation — Conflicting with the guideline in item 1, Material’s navigation organizes content hierarchically. Also to overcome the guideline described in item 5Material suggests to show content overflow in the view’s initial scroll position by cutting of content above the fold, encouraging further exploration, as illustrated in Fig. 4.18. Figure 4.18: Scrolling Guideline 17 •Icons — Material provides over 900 icons to be used in web, Android and iOS development 18. They can be defined as somewhat abstract graphics, going against the guideline in item 1. On the other hand these icons have been widely adopted, providing consistency between different applications. Also, in some components, under certain conditions, Material indicates that icons should not be accompanied by text labels, conflicting with the guideline presented in item 2. Finally, Material’s icons are not culturally sensitive, opposing item 3. 16https://material.io/guidelines/growth-communications/gesture-education.html 17https://material.io/guidelines/components/grid-lists.html#grid-lists-behavior 18https://material.io/icons/ 50 Methodology 4.4 Phase Four - Prototyping and Evaluation After looking into some design guidance from different authors, in section 4.3, it is time to develop concrete designs and evaluate them. 4.4.1 Application’s Navigation In order to design the navigation for the mobile application, common user tasks based on the system’s requirements (mentioned in section 4.2) were first prioritized, as illustrated in 4.19. Figure 4.19: Prioritisation of common user tasks The different priority levels helped to inform the application’s navigation, giving prominence in the UI to the tasks with highest priority. 4.4.2 Low Fidelity Prototype Low fidelity prototypes, as stated in 3.4.4.1 provide a quick and easy way to create alternative design versions. Firstly, sketches on paper (see Fig. 4.20) allowed to experiment with different layouts and to test possible application flows, taking into account requirements and tasks priorities. 51 Methodology Figure 4.20: Paper Sketches After some iteration the final low fidelity prototype was created using the Balsamiq Mockups 19 wireframing tool. The prototype’s screens are provided along with the application flow in Figures 4.21,4.22,4.23,4.24 and 4.25. ASTART START A0 LAUNCH SCREEN A1 LOGIN SCREEN A2 REGISTER SCREEN A3 INTRODUCTION SCREEN B0 PESTS VIEW A0 LOADING SCREEN Screen Name and ID Number Automatic Transition User Touch Transition NOTATION Figure 4.21: App Flow - Start 19https://balsamiq.com/ 52 Methodology BHOMEPAGE B0 PESTS VIEW B0.1 DISEASES VIEW B0.2 ADD DISEASES B1 MOTH VIEW B2 DRAWER MENU B2.0 MAP VIEW B2.1 REFERENCE VIEW B2.4 SENSORS VIEW B2.3 WEATHER VIEW B2.2 ALERTS VIEW ALERT NOTIFICATION B3 ALERT VIEW B3.1 Long press needed to select list item B2.2.1 DELETE ALERTS VIEW A0 LOADING SCREEN Screen Name and ID Number Automatic Transition User Touch Transition NOTATION Figure 4.22: App Flow - Homepage 53 Methodology Screens Changes Navigation — The navigation drawer was replaced by bottom navigation in order to provide quicker access to high priority tasks. This also meant that those functionalities would be available without the need to click on the hamburger icon, which raised concerns about being a recognizable action without a label attached. The expert functionality went from the Moth View (Screen B1 in Fig. 4.22) to also be included in the bottom navigation, allowing quick navigation between top-level views. Also, instead of having archived Alerts in the drawer or in the bottom navigation they can be accessed in the Moth View with a card displaying the latest alerts and a status of already read or not. Map View — The icon that opens a menu to change the map view (satellite, terrain or heatmap) was changed from the app bar to the top-left of the map viewing area. This was done because of the concern that the iconography was not clear enough for users to understand it would allow to perform actions related with the map. Insert Plot Info — In this interface the stepper were discarded because of the lack of information input needed. 60 Methodology Plot View — This view was redesigned to have cards nested inside cards, even though it went against Material’s guidelines. This allowed the user to perform the same actions with fewer clicks. Merge of views — The weather and sensors views were merged because of how related they were to one another, providing the user both data in a single view. Reference View — Previously organized with a collection of cards the reference view was altered to provide navigation between the content in the form of chapters. Table 4.6: Design Changes 4.4.4 Evaluation After creating the hi-fi prototype, it was time to evaluate it using real farmers and to do so a usability test was conducted with winegrowers from the DDR (the usability test protocol can be seen in detail in Appendix I). The prototyping tool Invision 26 and the Craft plugin 27 were used to 26https://www.invisionapp.com/ 27https://www.invisionapp.com/craft 61 Methodology add animations, gestures, and transitions to the static interfaces created using Sketch, as illustrated in Fig. 4.26. Figure 4.26: Craft in Sketch 4.4.4.1 Participants The number of required participants for the test was established as at least 5 and up to 10 people, as recommended by Nielsen’s and Landauer’s analysis of return on investment when performing a usability study [NL93]. The recruitment process was performed by a local cooperative in Peso da Régua. In order to get a good sampling, the selection criteria for the recruitment was based on three factors, the type of producer, the participant’s technological proficiency and educational level. It was important to perform tests with small and also medium/large winegrowers, with people that had never used a smartphone and others that owned one, and with very different levels of education. 4.4.4.2 Context of Use The test was conducted in a room in the cooperative facilities, a place familiar to all participants, which would allow them to be more comfortable. It is important to note that using the application in a room sitting down it is very different from the expected context of use. To even out this 62 Methodology difference the test was conducted next to a big source of light, in order to simulate the expected glare the users would get using the device outdoors. 4.4.4.3 Test procedure At first the participants were informed about the context of the study, similarly to the introduction given when performing the interviews (see Appendix D). Next it was emphasized that the test did not serve to evaluated the user, but the system. This was important because most of the users did not know what a usability test was. The test had a time limit of 1 hour and was conducted with only one person with the role of guiding the participant, before (trainer) and during the session (facilitator). In order to facilitate the recording of the participant’s actions and comments a tripod and camera were used to record the test with the consent of the participant. Participants were instructed to use the think-out-loud method during the test. After, if there was any free time remaining, an open conversation with the participant allowed to not only talk about aspects relating to the prototype’s usability but to possible improvements when it came to features. 4.4.4.4 Tasks As previously mentioned the hi-fi prototype was developed with the tasks that would later be evaluated in mind. The selection of those tasks was done based on two factors: the most important things every user would be able to accomplish using the app and also different types of interaction. For example, tasks such as creating and mapping a sample, trap or pheromone dispenser followed the same premise, therefore there was no need to evaluate all three of them. The tasks the users were asked to perform and the ideal flow for their successful completion is listed in table 4.7. Task Ideal Flow Task 1.1 — Change the data range of the moth’s flight curve to the 16th of April - 11th of June [Click 28th May button > Edit date in picker to 16th of May] Task 1.2 — See the number of captured adults and nests counted during the week of 23-29 April [Click on 23rd-29th of April] Task 2 — Open the latest alert sent out by DRAP [Click on button with label “Circular no 13” on “Avisos Agricolas” Card] Task 3 — See the scheduled treatments for the 10th and 11th of June and see which one has more urgency [Scroll to bottom of “Pagina Inicial” screen > Swipe left over dates on “Tratamentos” Card or Click on right arrow] 63 Methodology Task 4.1 —- Create and map a new plot [Click on “Mapa” in the bottom navigation > Click on plus FAB > Simulate click on screen to map points > Click on check FAB > Simulate field inputs > Click next > Simulate field inputs > Click on save floating action button] Task 4.2 —- Change the view to a heat map in order to see what areas of the plot are more prone to being attacked by the grapevine moth [Click on layers icon > Select on menu “Mapa de Temperatura”] Task 4.3.1. —- On that same plot, add and map a trap [Click on plot > Click on “Armadilhas” in “Prevencao” card > Click plus FAB > Simulate click on map to pin trap location > Simulate input of trap’s name > Click check FAB] Task 4.3.2. —- Now perform an automatic count of adult moths in that trap [Select day in date picker > Click Automatic Count > Simulate photo capture of trap > Click next > Input grey area manual count > Click Save] Task 5 —- Contact expert “Eng. Sandra” by chat and send her three images from sample called “Amostra #12” with a question attached [Click on “Mapa” in the bottom navigation > Click on plus FAB > Simulate click on screen to map points > Click on check FAB > Simulate field inputs > Click next > Simulate field inputs > Click on save floating action button] Task 6 - Read air humidity from the sensor in the weather station “Estacao 2” [Click on “Sensores” in the bottom navigation > Select weather station “Estacao 2”] Task 7 - Open reference information about how to apply traps [Click on “Referencia” in the bottom navigation > Search “armadilha” and go through search results OR Click on dropdown and go to “Estimativa de Risco”] Table 4.7: Usability Test Tasks 4.4.4.5 Metrics To measure the effectiveness of the system the following metrics were recorded: •Task completion rate — the task is considered completed when the task goal is achieved. •Errors — are considered errors actions that do not contribute to task completion. •Assistance — when the participants cannot proceed on a task, the test facilitator can give help. Two types of assists can be given, a less direct assistance (light assist), and a more 64 Methodology direct and intrusive assistance (heavy assist), for example the facilitator clicking on the screen for the participant, used only as a last resort. To evaluate the system’s efficiency the time on task (TOT) was recorded. Because the thinkout-loud method was used, for better comparison, the TOT measures only to the time the user interacted with the prototype, not the time when they conversed with the facilitator. Also satisfaction was measured through the System Usability Scale (SUS) questionnaire administered after the test. 4.5 Summary The user centred design methodology was used encompassing the following phases: •Phase 1: Requirements Gathering — During this phase an analysis of the existing android applications in the market were analysed, the research sample was established as winegrowers in the Douro region and research methods were used to collect data to inform the system’s requirements. Interviews and field visits were the chosen methods. •Phase 2: Analysis — After collecting data it was necessary to analyse it, therefore personas and problem and activity scenarios were created. Also use cases and hierarchical task analysis diagrams were made. The scope of the study evolved from Agriculture in general to Prevention and Control of the grapevine moth. •Phase 3: Design — With the requirements for the system established it was necessary to study design guidelines focused on the target audience and context of use. These would later inform the creation of the concrete designs. •Phase 4: Prototyping and Evaluation — Low and high-fidelity prototypes were designed and evaluated with real winegrowers by performing a usability test. The results from the usability test are presented and discussed in the next chapter 5. 65 Methodology 66 Chapter 5 Results and Discussion 5.1 Participants User # Gender Age ISCED Mobile Phone Type of producer U01 M 77 1 Has never owned one Small U02 F 64 3 Basic Phone Small U03 M 59 3 Smartphone Medium/Large U04 M 53 2 Smartphone Small U05 M 43 3 Smartphone Medium/Large U06 M 46 3 Smartphone Medium/Large Table 5.1: Participants As described in table 5.1 six participants were apart of this test, with an average age (rounded to nearest year) of 57 years old. Four out of six participants owned a smartphone, one running Apple’s operating system iOS and the rest android. Relatively to the other two participants, one had never owned a mobile phone, only using a fixed landline and the other participant used a basic phone, without a touch screen. Regarding the type of producer, three participants were small winegrowers and three medium/large winegrowers. Based on the International Standard Classification of Education (ISCED), one participant only finished primary education, one lower secondary education and four upper secondary education. Photos of participants performing the test can be seen in Fig. 5.1. 67 Results and Discussion Figure 5.1: Participants during the test 5.2 Performance results A full report on the metrics relating to efficiency and efficacy of the system can be read in Appendix J. A summary of the results is shown in table 5.2, which includes data relating to time on task, effectiveness of tasks, errors and assists. In Fig. 5.2 we can see task completion rate of each user, where it is possible to visualize that the first two users (U01 and U02), that were not smartphone users, were the ones who needed more assistance. In Fig. 5.3 data regarding the number of errors and assists by each user is provided. U01 had the most number of errors and assistances, probably due to his age and never having owned a mobile phone. Other relevant graphs that were created are: the one relating to user metrics by task, in Fig. 5.4. It is possible to conclude that the tasks that generated more issues were T3, T4.3.1 and T7 and the ones that were executed by most users were T1.2, T2, T4.3.2, and T6. Also the graph relating to the total analysis of the metrics by task, can be seen in Fig. 5.5. USER # TOTAL TIME ON TASK [SECONDS] TOTAL UNASSISTED TASK EFFECTIVENESS [(%)COMPLETE] TOTAL LIGHTLY ASSISTED TASK EFFECTIVENESS [(%)COMPLETE] TOTAL HEAVILY ASSISTED TASK EFFECTIVENESS [(%)COMPLETE] TOTAL ERRORS TOTAL ASSISTS (LIGHT) TOTAL ASSISTS (HEAVY) U01 1142 0% 46% 100% 26 26 7 U02 674 28% 55% 100% 11 12 6 U03 141 91% 100% 100% 2 1 0 U04 303 82% 91% 100% 2 2 1 U05 466 64% 82% 100% 11 5 2 U06 469 64% 73% 100% 0 6 3 Mean 532,5 55% 80% 100% 8,67 8,67 3,17 Standard Deviation 349 40% 30% 0% 9,8 9,4 2,8 Min 141 0% 46% 100% 0 1 0 Max 1142 91% 100% 100% 26 26 7 Table 5.2: Summary of results 68 Results and Discussion Figure 5.2: Task completion rates Figure 5.3: Total errors and assists 69 Results and Discussion Issue Recommendation Scrolling — After conducting the test it was possible to see that some of the guidelines mentioned in section 4.3 were quite helpful to inform the design of the hi-fi prototype but they were not all strictly followed and because of this some of those issues still occasionally arose. This is an example of that, some participants had a hard time realising that scrolling was possible even following the scrolling guideline provided by Material. This happened because the device used while develop the design and the device used during the tests had different screen resolutions. To combat this, providing a scrollbar on the screen could help to inform the users scrolling was possible or making sure, that with different screens resolutions the guideline of showing content overflow would be effective. Buttons — For the less experienced users clicking smaller buttons or buttons with a smaller or not defined touch area proved to be hard. If not constricted by other element on the UI, button’s touch area should be larger than the Material’s default area. Mapping plot — Some users thought the creation of parcel would be done by pressing points on the map directly, without having yo press the plus FAB first. This was actually something that was taken into account when designing. In order to prevent the user clicking on a plot while trying to map a new one, it was necessary to add a button to allow the user to map a new plot. A probable reason for this issue was that either people did not realise a FAB add button was available (due to it being green and the map behind having a majority of green tones) or they did not associate the symbolism of the plus icon with mapping a plot. A way to overcome this would be to change the FAB colour to the secondary colour (orange) for it to stand out and/or to change the icon to a pencil or even add a label to it. This should be applied to other interfaces in order to provide consistency. Treatments’ Calendar — Some users mixed up the treatment’s starting date with the day the treatment was scheduled to be applied. In order to avoid this mix-up the treatment’s starting date should be omitted from the calendar view and only if the user opened the treatment this information would be provided. 76 Results and Discussion Icons — Some of Material’s icons proved to be too abstract for some users, especially the “layers” icon in the map view, even though it is also used in other popular Google mobile apps such as Google Maps. As mentioned in section 4.3 icons should be culturally sensitive and not very abstract. Therefore an activity, for example, card sorting could help to inform designers on what icons farmer easily recognize and the symbolism they provide. Heatmap — Most users were not familiar with the concept of a heatmap, therefore they assumed it meant to view the temperature that plot. To combat this misconception this functionality should be presented in the Introduction screen. Automatic Count — The grey area of uncertainty on the automatic count functionality was not easily understood by some users. Providing a help dialog on the first times the user used this functionality would allow to inform the user about what the grey area means and about what the user should do next. Reference Material — Most users had a hard time understanding where they could find information about trap’s placement This issue originated probably due to an incorrect task formulation. The task made the assumption that the farmer was used to deal with the grapevine moth and that he was also familiar with concepts related to prevention and control of pests. Besides this users had difficulties viewing the arrow on the dropdown button. This issue should be tackled and an improvement to this interface would be to have Material card stacked, simulating a book and allowing the user to swipe left/right or clicking on arrows to navigate through the different chapters. Another issue was that none of the users realised that the search option was available, this could possibly be overcome by providing a spotlight walkthrough on each main functionality for first time users. Table 5.5: Recommendations 5.6 Guidelines Taking into account all the information mentioned above here is a list of guidelines for designing for similar target-users. •“First time users” — Most farmers said that if they experimented or had the help from someone they would feel more confident using the system and that it would be a factor that would influence their technology acceptance. For first time users I would suggest providing 77 Results and Discussion a quickstart model 1. Also providing help dialogs could be beneficial for more complex tasks. •“Buttons” — Provide large touch target areas that are not nearby other touch areas. Delimited buttons seemed to be more recognizable than buttons without borders around. Avoid Material’s flat buttons, if needed, provide a simple thin black border around the button area. •“Gestures” — Provide buttons for multi-finger gestures because some users aren’t aware of those functionalities (e.g. zoom magnifier on map). In addition providing alternative ways to perform simple gestures, like swiping, using arrows should be provided in some cases. Also be aware that some users not that familiar with touch screens will perform long-presses involuntarily, so try to avoid that type of gesture. •“Colour” — It is important to, not only, take special attention to colour contrast when talking about elements in components, but also to pay attention to how colours on the interface compare to each other in terms of what elements are most prominent. For example, if using FABs, make sure that they are a colour that is very distinctive from other colours on the interface. •“Graphics” — The use of graphical symbols and photos seemed to translate very well with all users. If possible use pictorial representations to assist and encode information. •“Icons” — It is important to test with the target-users icons and the symbolism associated with them. •“Navigation” — Keep the app’s navigation simple by providing more sequential tasks. •“UI Components” — Material’s components seemed to not pose many problems for smartphone owners but for those who weren’t familiar with other apps using Material’s guidelines and components some issues arose. The components that should be avoided are the ones that require the user to be aware of actions that are not explicitly informed using the component. For example bottom navigation, if possible, should be preferred over a navigation drawer. 5.7 Summary From a usability test with six winegrowers, with different profiles, it was possible to come to conclusions relating technological acceptance and user needs. Different metrics relating to the system’s efficiency and efficacy were analysed and also relating to satisfaction. Commentaries and recommendations were also given based on this data. 1https://material.io/guidelines/growth-communications/onboarding.html# onboarding-quickstart 78 Chapter 6 Conclusions and Future Work 6.1 Contribution The aim of this study was to develop and adapt a mobile solution based on farmers’ needs and context of use. To do so an in-depth research was conducted following a user centred design methodology. This allowed to discover several areas of opportunity for ICT solutions for Viticulture in the Douro region. This study hopefully marks the beginning of more projects related to mAgri solutions for rural areas in Portugal and will help inform the design of those systems when it comes to farmer-computer interaction. 6.2 Challenges By far the main challenge in this research was the dependency on other people and organizations to provide knowledge, because the subject area of this study was so far off from my areas of study. Time and bureaucratic constraints made it difficult to follow an iterative development that is so crucial to the user centred design methodology. Looking back, having input from real farmers since the earlier stages would have been very beneficial. 6.3 Future Work Firstly, as previously said, redesigning the prototype according to the recommendations given and performing iterations of tests with farmers would be the future work developed. The natural step after that would be to implement and deploy the application and to do so it would be necessary to construct a prediction model for pests and diseases in Portugal’s vineyard. It would also be relevant to investigate on how computer vision solutions would allow to perform automatic count of insects in traps. 79 Conclusions and Future Work 80 References [Acc17] AccuWeather. Accuweather, 2017. https://play.google.com/store/ apps/details?id=com.accuweather.android&hl=en. [Agr16] Agrivi. Agrivi, version (2.4.1), 2016. https://play.google.com/store/ apps/developer?id=Agrivi&hl=en. [air15] Case Study: Airtel Kilimo, Kenya. Technical Report February, GSMA, 2015. [AMkP04] Chadia Abras, Diane Maloney-krichmar, and Jenny Preece. User-Centered Design. pages 1–14, 2004. [Ava13] Avallain Developer. ikilimo (version 1.0 ), 2013. https://play.google.com/ store/apps/details?id=com.avallain.ikilimo&hl=en. [Bar16] Barchart.com. Agmobile, version (1.2.3), 2016. https://play.google.com/ store/apps/details?id=com.agmobile.app&hl=en. [Ben10] David Benyon. Designing Interactive Systems: A Comprehensive Guide to HCI and Interaction Design (2nd Edition). Pearson Education Canada, 2010. [BLD04] R. Bongiovanni and J. Lowenberg-Deboer. Precision agriculture and sustainability. Precision Agriculture, 5(4):359–387, aug 2004. [BM05] Randolph G. Bias and Deborah J. Mayhew. Cost-Justifying Usability: An Update for the Internet Age, Second Edition (Interactive Technologies). Morgan Kaufmann, 2005. [Boe88] B. W. Boehm. A spiral model of software development and enhancement. Computer, 21(5):61–72, may 1988. [Bra] Terry Brase. Precision agriculture. http://www.delmarlearning.com/ companions/content/140188105X/trends/history_pre_agr.asp. (Accessed on 03/15/2017). [CA06] Cristina Carlos and Ana Aguiar. Estimativa do Risco de Traça da Uva na Região Demarcada do Douro. 2006. [Car12] Cristina Carlos. 14013537844428.pdf. http://www.advid.pt/imagens/ cadernos/14013537844428.pdf, 2012. (Accessed on 05/24/2017). [CLL15] M. Castle, B. D. Lubben, and J. Luck. Precision agriculture usage and big agriculture data | agricultural economics. http://agecon.unl.edu/cornhusker-economics/2015/ 81 REFERENCES precision-agriculture-usage-and-big-agriculture-data, May 2015. (Accessed on 02/06/2017). [CR87] John M. Carroll and Mary Beth Rosson. Interfacing Thought: Cognitive Aspects of Human-Computer Interaction (Bradford Books). The MIT Press, 1987. [Cra43] K. J. W. Craik. The Nature of Explanation. Cambridge University Press, 1943. [Dav89] Fred D. Davis. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3):319, sep 1989. [DFAB03] Alan Dix, Janet E. Finlay, Gregory D. Abowd, and Russell Beale. Human-Computer Interaction (3rd Edition). Pearson, 2003. [DM97] Peter J. Denning and Robert M. Metcalfe. Beyond Calculation: The Next Fifty Years of Computing. Copernicus, 1997. [DMG+] Tiago Devezas, Job Mashapa, Bruno Giesteira, Darelle Greunen, and Carlos Carreira. ICT4D HCI Guidelines : A study for Developing Countries Research-based HCI Guidelines for developing countries. (2014):189–205. [Eas87] K. Eason. Information technology and organizational change, 1987. [Esp16] Espiralpixel. Openpd (version 0.0.15 ), 2016. https://play.google.com/ store/apps/details?id=com.ionicframework.OpenPD678289. [Fit54] Paul M. Fitts. The information capacity of the human motor system in controlling the amplitude of movement. Journal of Experimental Psychology, 47(6):381–391, 1954. [Gie15] Bruno Giesteira. HCI4D Guideline Systematization: Creation, Documentation and Evaluation with Partners from Developing Countries. 26(1):1114–1128, 2015. [GKM12] Elizabeth Goodman, Mike Kuniavsky, and Andrea Moed. Observing the User Experience: A Practitioner’s Guide to User Research. Morgan Kaufmann, 2012. [GL85] John D. Gould and Clayton Lewis. Designing for usability: key principles and what designers think. Communications of the ACM, 28(3):300–311, March 1985. [GWO+16] A. Gichamba, P. Waiganjo, D. Orwa, R. Wario, and B. Ngari. Prototyping magriculture applications among smallholder farmers. In 2016 IST-Africa Week Conference, pages 1–7, May 2016. [HGWH13] Tim Hatt, Corina Gardner, Adams Wills, and Martin Harris. Scaling Mobile for Development. Technical Report August, GSMA Mobile for Development Intelligence, 2013. [HH89] H.Rex Hartson and Deborah Hix. Toward empirically derived methodologies and tools for human-computer interface development. International Journal of ManMachine Studies, 31(4):477–494, oct 1989. [HJ93] Karen Holtzblatt and Sandra Jones. Participatory Design: Principles and Practices. CRC Press, 1993. [INE13] Inquérito à Estrutura das Explorações Agrícolas 2013. Instituto Nacional de Estatística, edição 201 edition, 2013. 82 REFERENCES [INE15] Estatísticas Agrícolas 2015. Instituto Nacional de Estatística, 2015. [Int15a] Gsma Intelligence. Agricultural machine-to-machine (Agri M2M): a platform for expansion. (March), 2015. [Int15b] Gsma Intelligence. Agricultural value-added services (Agri VAS): market opportunity and emerging business models. (February), 2015. [Ira10] Marie Louise Iraba. Inexpensive Mobile Technologies to Empower Rural Farmer with M-Agriculture. PhD thesis, University of the Western Cape, 2010. [JH02] Jeff Johnson and Austin Henderson. Conceptual models: begin by designing what to design. interactions, 9(1), jan 2002. [Kar94] Clare-Marie Karat. Usability Inspection Methods: A comparison of user interface evaluation methods. Wiley, 1994. [KL15] Sami Khanal and Rattan Lal. Precision agriculture for improving water quality under changing climate. In Advances in Soil Science, pages 283–306. Informa UK Limited, jul 2015. [KS98] Gerald Kotonya and Ian Sommerville. Requirements Engineering: Processes and Techniques. Wiley, 1998. [Lan95] Thomas K. Landauer. The Trouble with Computers: Usefulness, Usability, and Productivity. The MIT Press, 1995. [Low13] Travis Lowdermilk. User-Centered Design: A Developer’s Guide to Building UserFriendly Applications. O’Reilly Media, 2013. [M+16] Ville Myllynpää et al. Mobile applications, solution for sustainable agriculture?-study of magriculture services in kenya. 2016. [Nie94] Jakob Nielsen. Usability Engineering (Interactive Technologies). Morgan Kaufmann, 1994. [Nie95] Jakob Nielsen. 10 heuristics for user interface design. https://www. nngroup.com/articles/ten-usability-heuristics/, 1995. (Accessed on 02/05/2017). [NL93] Jakob Nielsen and Thomas K. Landauer. A mathematical model of the finding of usability problems. In Proceedings of the SIGCHI conference on Human factors in computing systems - CHI 93. ACM Press, 1993. [NN] Don Norman and Jakob Nielsen. The definition of user experience (ux). https: //www.nngroup.com/articles/definition-user-experience/. (Accessed on 02/06/2017). [Nor93] Donald A. Norman. Things That Make Us Smart: Defending Human Attributes In The Age Of The Machine (William Patrick Book). Basic Books, 1993. [Nor11] Donald A. Norman. The Design of Everyday Things. Tantor Audio, 2011. [Nor13] Don Norman. The Design of Everyday Things: Revised and Expanded Edition. Basic Books, 2013. 83 REFERENCES [OT11] Shannon O’Day and Louis Tijerina. Legibility: Back to the basics. SAE Int. J. Passeng. Cars – Mech. Syst., 4:591–604, 04 2011. [PCPC13] Emanuele Pierpaoli, Giacomo Carli, Erika Pignatti, and Maurizio Canavari. Drivers of precision agriculture technologies adoption: A literature review. Procedia Technology, 8:61–69, 2013. [PEA17] PEAT. Plantix (version 1.2.0), 2017. https://play.google.com/store/ apps/details?id=com.peat.GartenBank&hl=en-ca. [Phi11] Tanid Phiriyapokanon. Is a big button interface enough for elderly users ? 2011. [Pro] Prototyping | usability.gov. https://www.usability.gov/ how-to-and-tools/methods/prototyping.html. (Accessed on 02/05/2017). [PSLDS97] F.J. Pierce, E.J. Sadler, J. Lowenberg-DeBoer, and S.M. Swinton. Economics of sitespecific management in agronomic crops. In The State of Site-Specific Management for Agriculture. American Society of Agronomy, 1997. [PSR15] Jenny Preece, Helen Sharp, and Yvonne Rogers. Interaction Design: Beyond HumanComputer Interaction. Wiley, 2015. [RB88] W.A.S Baecker R.M. Buxton. Readings in Human-Computer Interaction: A Multidisciplinary Approach. Morgan Kaufmann Publishers In, 1988. [RC01] Mary Beth Rosson and John M. Carroll. Usability Engineering: Scenario-Based Development of Human-Computer Interaction (Interactive Technologies). Morgan Kaufmann, 2001. [RC12] Mary Beth Rosson and John M. Carroll. Human Computer Interaction Handbook: Fundamentals, Evolving Technologies, and Emerging Applications, Third Edition (Human Factors and Ergonomics). CRC Press, 2012. [Saf09] Dan Saffer. Designing for Interaction: Creating Innovative Applications and Devices (Voices That Matter). New Riders, 2009. [SLD98] S. M. Swinton and J. Lowenberg-DeBoer. Evaluating the profitability of site-specific farming. jpa, 11(4):439, 1998. [SP03] Ben Shneiderman and Catherine Plaisant. Designing the User Interface (Fourth Edition Preview, Third Edition). Addison Wesley, 2003. [T+16] Daniele Tricarico et al. Agricultural Value-added Services (Agri VAS) Toolkit 2.0. Technical Report April, GSMA, 2016. [Tse15] Lilian Tse. Unlocking the rural market with user centered design. http: //www.gsma.com/mobilefordevelopment/wp-content/uploads/ 2015/11/LilianTse_Frog.pdf, October 2015. (Accessed on 03/27/2017). [Wes14] Michael Westcott. Design management institute. http://www.dmi.org/ blogpost/1093220/182956/, March 2014. (Accessed on 03/27/2017). [WEU12] Gary Woodill, D Ed, and Chad Udell. mAgriculture : The Application of Mobile Computing to the Business of Farming. Technical report, float, 2012. 84 REFERENCES [WSC+16] Susan Wyche, Charles Steinfield, Tian Cai, Nightingale Simiyu, and Martha E Othieno. Reflecting on Video : Exploring the Efficacy of Video for Teaching Device Literacy in Rural Kenya. (January), 2016. [Z+15] Qin Zhang et al. Precision Agriculture Technology for Crop Farming. CRC Press, 2015. 85 - - referenced in papers price Free Various(2) location Various(1) (1) (2) Plantwise Factsheets Library by CABI last update 10/19/16 rating 4.34 44 installs 1+ thousand top ranking age February 2014 categories languages price Free English location - Doodo by Doodo technology inc. last update 12/26/16 rating 4.45 22 installs 100+ top ranking age November 2016 categories languages price Free English Gujarati Hindi Marathi location India AgroStar - Agriculture App by Ulink Agri-Tech Pvt. Ltd. last update 2/23/17 rating 4.36 949 installs 50+ thousand top ranking age March 2016 categories languages Diseases and Pests Apps Diseases and Pests Apps Afghanistan, Bangladesh, Barbados, Bolivia, Brazil, Burkina Faso, Cambodia, China, Democratic Republic of Congo, Costa Rica, Ethiopia, Ghana, Grenada, Honduras, India, Jamaica, Kenya, Malawi, Mozambique, Myanmar, Nepal, Nicaragua, Pakistan, Peru, Rwanda Sierra Leone, Sri Lanka, Tanzania, Thailand, Trinidad and Tobago, Uganda, Vietnam, Zambia Bengali, Chichewa, Chinese, Dari, English, French, Khmer, Kinyarwanda, Nepali, Portuguese, Quechua, Sinhalese, Spanish, Swahili, TamilThai Agriculture Information Apps India - - price Free English Kannada location India Way2ABI Agri FBI by Way2Agribusiness India Pvt Ltd last update 11/14/16 rating 4.60 42 installs 1+ thousand top ranking age November 2015 categories languages price Free English location Zimbabwe iFarm Mobile by iFarm Zimbabwe (Private) Limited last update 2/19/17 rating 4.20 5 installs 100+ top ranking age November 2016 categories languages Market Data Apps Business Apps Weather Apps by top ranking last update rating pricelocation installs categories languages age December 2015 -English 5 1 Ag Guardian Ag Guardian, LLC 1/1/16 Free100+ Business Apps referenced in papers referenced in papers referenced in papers by top ranking last update rating pricelocation installs categories languages age December 2013 USA Brazil France Spain Germany Indonesia India Mexico Switzerland Netherlands Various(1) 4.42 27044 GPS Fields Area Measure Studio Noframe 9/16/16 Free1+ million by top ranking last update rating price installs categories languages age November 2013 -English 3 53 AgDNA AgDNA 6/24/16 Free location USA*, Canada* Australia* New Zealand* 1+ thousand by top ranking last update rating pricelocation installs categories languages age August 2012 -English 4.2 325 FarmLogs AgriSight Inc. 3/8/17 FreeUSA* 10+ thousand (1) Afrikaans, Arabic, Bulgarian, Catalan, Chinese (simplified), Croatian, Czech, Dutch, English, Finnish, French, Galician, German, Greek, Hebrew, Hungarian, Indonesian, Italian, Korean, Latvian, Lithuanian, Norwegian (Bokmål), Norwegian (Nynorsk), Persian, Polish, Portuguese, Portuguese (Brazil), Romanian, Russian, Serbian, Slovak, Slovenian, Spanish, Spanish (Argentina), Spanish (Colombia), Spanish (Mexico), Swedish, Tamil, Thai, Turkish, Ukrainian, Urdu Field Mapping Apps Field Mapping Apps Field Mapping Apps referenced in papers by top ranking last update rating pricelocation installs categories languages age January 2012 France Germany Spain Netherlands Russia Switzerland Brazil English 4.41 868 eFarmer eFarmer B.V. 2/23/17 Free + In-AppNetherlands* 50+ thousand by top ranking last update rating pricelocation installs categories languages age October 2014 Saudi Arabia Indonesia Netherlands Switzerland English 4.41 176 AgMobile Barchart.com 2/26/16 Free + In-AppUSA* 10+ thousand by top ranking last update rating pricelocation installs categories languages age November 2013 GermanyGerman 4.4 502 BayWa Agri-Check BayWa AG 2/28/17 FreeGermany* 50+ thousand Farm Management Apps Field Mapping Apps Weather Apps Market Data Apps Agriculture Information Apps Weather Apps Market Data Apps referenced in papers by top ranking last update rating pricelocation installs categories Farm Management Apps languages age January 2016 -English 0 0 AgVault Sentera 1/26/16 FreeUSA* 100+ by top ranking last update rating pricelocation installs categories languages age May 2014 India English Hindi Bengali Gujarati Kannada Marathi Punjabi Tamil Telugu 4.32 7606 RML Farmer - Krishi Mitr RML AgTech 2/2/17 FreeIndia 100+ thousand by top ranking last update rating pricelocation installs categories languages age March 2016 IndiaHindi Gujarati 4.54 751 AGRISCIENCE KRISHI Infinium Global Solutions 3/23/17 FreeIndia 50+ thousand Weather Apps Market Data Apps Diseases and Pests Apps Learning and Reference Apps Market Data Apps Learning and Reference Apps by top ranking last update rating pricelocation installs categories languages age October 2015 IndiaMarathi English Hindi 4.38 1030 KrushiKing  Krushiking Agrotech Industries Pvt Ltd. 3/14/17 FreeIndia 10+ thousand by top ranking last update rating price installs categories languages age February 2017 -English 0 0 Huron Tractor AgDNA 3/4/17 Free 5+ by top ranking last update rating pricelocation installs categories languages age January 2013 -English 4.07 103 Bayer TurfXpert 2 dam creative 3/2/17 FreeUK 10+ thousand Learning and Reference Apps Weather Apps Market Data Apps Agriculture Information Apps location USA*, Canada* Australia* New Zealand* Farm Management Apps Field Mapping Apps Weather Apps Agriculture Information Apps Farm Management Apps Field Mapping Apps Weather Apps Diseases and Pests Apps referenced in papers by top ranking last update rating pricelocation installs categories languages age October 2016 IndiaGujarati 4.7 1896 Khedut Mitra - GJ Agriculture FireTruck Enter10ment 3/21/17 FreeIndia 50+ thousand by top ranking last update rating pricelocation installs categories languages age September 2015 India English Punjabi Malayalam Bengali Odia Marathi Kannada Telugu Tamil Gujarati Hindi 4.19 2070 IFFCO KisanAgriculture App IFFCO Kisan 1/3/17 FreeIndia 100+ thousand by top ranking last update rating pricelocation installs categories languages age April 2016 -English Hindi 4.88 436 Agribolo - Agriculture App Agrilife Technologies Pvt. Ltd. 2/7/17 FreeIndia 5+ thousand Market Data Apps Business Apps Weather Apps Market Data Apps Agriculture Information Apps Learning and Reference Apps Learning and Reference Apps Weather Apps Market Data Apps Agriculture Information Apps FINDINGS market analysis of android apps (that appear more than once) (that appear more than once) TOP 10 (Rating)* APP NAME RATING NUMBER OF RATINGS Agribolo - Agriculture App 4.88 436 Khedut Mitra - GJ Agriculture 4.7 1896 AGRISCIENCE KRISHI 4.54 751 agrofarm 4.47 159 GPS Fields Area Measure 4.42 27044 eFarmer 4.41 868 AgMobile 4.41 176 BayWa Agri-Check 4.4 502 KrushiKing   4.38 1030 Plantix 4.38 322 * with more than 100 ratings 4.1.2 Locations4.1.1. Categories 4.1.3. Languages 4.1.4. Age 4.1.5. Monolingual VS Multilingual TOP 10 (Installs) APP NAME NUMBER OF INSTALLS GPS Fields Area Measure 1+ million RML Farmer - Krishi Mitr 100+ thousand AgriApp 100+ thousand IFFCO KisanAgriculture App 100+ thousand AGRISCIENCE KRISHI 50+ thousand eFarmer 50+ thousand BayWa Agri-Check 50+ thousand Plantix 50+ thousand AgroStar - Agriculture App 50+ thousand Khedut Mitra - GJ Agriculture 50+ thousand 4.1.6. 4.1.7. Most Versatile Apps APP NAME NUMBER OF CATEGORIES IFFCO KisanAgriculture App 5 Agribolo - Agriculture App 4 Bayer TurfXpert 4 Huron Tractor 4 KrushiKing  4 RML Farmer - Krishi Mitr 4 4.1.8. Market Analysis 100 Appendix B Categorizing Activity B.1 Introduction After collecting data about several apps released in the play store market it was necessary to assign categories to each app, depending on the functionalities provided or marketed by the developer. The “mother-categories” where defined by a previous study developed by Gary Woodill and Chad Udell, dating back to 2012. These study listed 8 different categories for mobile agriculture apps: •Information Apps — These applications provide general information about agriculture, such as, laws and regulations, as well as information portals. •Business Apps — Dedicated to the business aspect of the industry, they allow you to calculate investment estimates for materials or changes to the terrain. •Diseases and Pests Apps — This category is intended for the prevention and control of diseases and pests. They can be informative as well as having the ability to recognize a disease through a photograph of the leaf. •Farm Management Apps — They help with better management, such as personnel management, resources, etc. •Field Mapping Apps — Through the GPS functionality it is possible to make a satellite mapping of the terrain, in order to measure and identify terrain areas. •Learning and Reference Apps — They serve to educate and to be used as reference. They include dictionaries, encyclopedias, courses, etc. •Market Data Apps — They allow access to information about the market, such as price of various products, sometimes updated in real time. •Weather Apps — Dedicated to weather monitoring, they sometimes allow you to receive alerts of bad weather or natural phenomena. 101 Statistical Data - Portugal and Douro Region Figure C.2: Employment rate, persons aged 20-64 •Working age population: Figure C.3: Working age population 108 Statistical Data - Portugal and Douro Region C.2 Education •Literacy: Definition: age 15 and over can read and write Total population 95.7% Male 97.1% Female 94.4% Table C.1: Literacy Rate 2 Figure C.4: Iliteracy Rate 3 2https://www.cia.gov/library/publications/the-world-factbook/geos/po.html 3https://www.ine.pt 109 Statistical Data - Portugal and Douro Region •Share of early leavers from education: “Share of young people aged 18–24 who were early leavers from education and training, by NUTS 2 regions, 2015 (%).” Figure C.5: Share of early leavers from education •Share of persons with tertiary education: “Share of persons aged 30–34 with tertiary education (ISCED levels 5–8) attainment, by NUTS 2 regions, 2015 (%)” 110 Statistical Data - Portugal and Douro Region Figure C.6: Share of persons with tertiary education C.3 Information Society •Broadband availability: Figure C.7: Proportion of households with broadband connections 111 Statistical Data - Portugal and Douro Region •Internet usage: Figure C.8: Proportion of people who never used the internet All this data was collected from The World Factbook by the Central Intelligence Agency4, from Eurostat5and from Statistics Portugal. 4https://www.cia.gov/library/publications/the-world-factbook/geos/po.html 5http://ec.europa.eu/eurostat/statistics-explained/index.php/Eurostat_regional_ yearbook 5https://www.ine.pt 112 Appendix D Script of Interview D.1 Introduction Hi my name is Maria Marques and I am a student at the Faculty of Engineering at the University of Porto. I am currently developing my master thesis at Fraunhofer Portugal. Fraunhofer is a private non-profit association founded by Germany’s largest applied research organization. I’ve also come accompanied by Eduardo Pereira, who is a researcher in the Human Computer Interaction field at Fraunhofer. Fraunhofer Portugal is dedicated to researching how to improve the living conditions of people through technological solutions capable of facilitating access to information and communication technologies. There are two specific groups that are the focus of that research, elderly people and populations in remote areas or in developing countries. In this case, the focus are farmers in rural areas of Portugal. We intend to increase their productivity and production through the development of a mobile application for smartphones. In order to better pay attention to you, I would like to ask for your permission to record audio. It is purely for the purpose of research and will not be used for commercial or promotional purposes. And I will also ask you to fill out a short questionnaire [deliver questionnaire in ??]. I would like you to read and sign the consent [deliver the consent form in D.4]. Do you have any questions before we start? D.2 Warm-up •Ask the participants to introduce themselves. What is your name, age, and what do you do? •How is a daily routine? Who do they work with and deal with directly? •Who do they trust/influences their decision making? And who is dependant on them to make decisions? •Do you own a phone? What type of phone? What do they use it for? 113 Script of Interview D.3 General questions In order to increase the productivity and production of the farmer in rural areas and focusing on the factors that lead to a good harvest: in grape quality and quantity as well as in the prevention and control of diseases and pests. D.3.1 For Farmers •What is your work process? Before, during and after harvesting. •Regarding the forecasting of production, what is your role and what is not performed by you? •Regarding the quality of the grape, what is your role and what is not performed by you? •How do you inform yourself of pests or diseases in the area? And about treatment? •What are your main goals? Optimize production? Better quality grapes? •What frustrates or discourages you? •Do you imagine yourself using a mobile phone on a daily basis in the field? D.3.2 Other stakeholders •In your opinion what are the farmers’ main goals within these themes? •And the main problem areas or concerns? •How could your work be facilitated using a mobile app? •How could a smartphone be integrated into the daily routine of a farmer? 114  " CONSENTIMENTO"PARA"PARTICIPAÇÃO"EM"INVESTIGAÇÃO" No"âmbito"do"projeto"«EyesOnFarm»,"estamos"a"realizar"um"estudo"com"o"objeIvo"de"desenvolver" uma"solução"para"disposiIvos"móveis,"focada"em"agricultores"da"vinha,"na"zona"do"Douro."Com"essa" solução" pretendemos" aumentar" a" produIvidade" e" produção" dos" agricultores," através" de" uma" invesIgação"em"que"estes"se"encontram"no"centro"do"processo"de"desenvolvimento."" Para"efeitos"de"pesquisa,"iremos"proceder"à"recolha"de"dados"sociodemográficos"e"à"gravação"audio" desta"entrevista." Gostaríamos"de"contar"com"a"sua"parIcipação."A"parIcipação"não"envolve"qualquer"prejuízo"ou"dano" material" e" não" haverá" lugar" a" qualquer" pagamento." Os" dados" recolhidos" são" confidenciais." A" Fraunhofer* AICOS* Portugal* tomará" todas" as" medidas" necessárias" à" salvaguarda" e" protecção" dos" dados"recolhidos"por"forma"a"evitar"que"venham"a"ser"acedidos"por"terceiros"não"autorizados." A" sua" parIcipação" é" voluntária," podendo" em" qualquer" altura" cessá-la" sem" qualquer" Ipo" de" consequência."Agradecemos"muito"o"seu"contributo,"fundamental"para"a"nossa"invesIgação!" O"parIcipante:" Declaro* ter* lido* e* compreendido* este* documento,* bem* como* as* informações* verbais* fornecidas* e* aceito*par@cipar*nesta*inves@gação.*Permito*a*u@lização*dos*dados*que*forneço*de*forma*voluntária,* confiando*que*apenas*serão*u@lizados*para*inves@gação*e*com*as*garan@as*de*confidencialidade*e* anonimato*que*me*são*dadas*pelo*inves@gador.*Autorizo*a*comunicação*de*dados*de*forma*anónima* a*outras*en@dades*que*estabeleçam*parceria*com*a*Fraunhofer*AICOS*Portugal*para*fins*académicos* e*de*inves@gação*cienIfica." Nome"do"parIcipante:"__________________________________________________________________" Assinatura"do"parIcipante:"______________________________________________________________" Data"___"/"___"/"______" InvesIgador"responsável:" Nome:"Eduardo"Pereira"""""""""Telefone:"+351"22"0430"363""""E-mail:"[email protected]" ESTE"DOCUMENTO"É"FEITO"EM"DUPLICADO:"UM"PARA"O"PARTICIPANTE"E"OUTRO"PARA"O" INVESTIGADOR." / 1 1 Script of Interview D.4 Consent Form 115 Script of Interview 116 Appendix E Pests and Diseases in Vineyard 117 Pests and Diseases in Vineyard 124 Appendix F Brainstorming Activity F.1 Introduction After interviewing and visiting various stakeholders in the winegrowing business it was time to assess the opportunities for development and decide which area the mobile solutions would tackle. F.2 Goal To come up with a user profile and ideas for a mobile android application relating to Viticulture in the Demarcated Douro Region. F.3 Participants The same participants of the Categorizing activity (Appendix B): •Participant 1 — Software engineer/Interaction designer •Participant 2 — Designer/Interaction designer •Participant 3 — Interaction designer with experience researching ICT4D solutions in under developed areas. F.4 Methodology The preparation of this activity was guided by the methods “Bundle Ideas” 1and “Brainstorming Rules” 2provided by IDEO. •Warm-up + Introduction — the participants were made aware of what phase in the design process this activity fit in. 1http://www.designkit.org/methods/28 2http://www.designkit.org/methods/30 125 Brainstorming Activity •Understanding the context — during this time mainly concepts and information related to viticulture and relevant to the activity were presented. •Presentation of personas and scenarios — in order to better present the findings from the research methods applied, scenarios and personas were presented, enabling the designers to get a better understanding of the possible users this solution would be designed for and their struggles. •Brainstorming activity — during this period of time discussion of ideas for application concepts and functionalities were discussed. A deck of post-its and a marker where delivered to each participant, motivating discussing and creativity and helping to better visualize the ideas that were generated. F.5 Conclusion Two themes for a possible mobile solution were taken into account, those being, Ripening Control and Pests and Diseases Prevention and Control. Also, two distinct personas were presented: the smaller, less educated, older farmer and the medium/larger winegrower, that is more used to interact with ICTs and more knowledgeable. Attached to each persona, a problem scenario was also shown. Read a more detailed description on the scenarios and personas created in sections ?? and 4.2.1.2. These are the ideas and conclusions that resulted from this activity: •Small winegrower — For the smaller winegrower persona, which probably has never even used a smartphone, a good idea to work around the lack of interaction with complex ICTs would be to have an informal caregiver, for example a son or younger member of the family to bridge the gap between the technology and the farmer. – Ripening Control — Related to Ripening Control an interesting idea to come up in this activity was to provide a less complex mobile solution, for example, using an USSD protocol for the user to check the laboratory results. – Pests and Diseases Prevention and Control — due to the complexity of this theme it would be hard to provide functionalities besides presenting reference material using a messaging system. •Medium/Large winegrower — This user profile allows for more complex functionalities and the integration with other technologies, and for higher interaction levels. – Ripening Control — Within this theme two distinct players are the laboratory technicians and the farmer. Some ideas presented were: for the laboratory were to: Send/- Make reports available, Optimize/computerise data input in the system, Automatic generation of reports with graphs and other visualization methods. For the farmer he would: Receive alerts and the reports and Visualize this data. A problem emerges, 126 Brainstorming Activity this mobile solution would require a lower degree of interaction from the farmer and would be mostly used from the laboratory side. As the goal of this dissertation is to study farmer’s interaction with a mobile solution this theme is not the most suitable. – Pests and Diseases Prevention and Control — This theme also has two players: the expert side, that being an association or a cooperative, and the farmer. For the cooperative some functionalities would be: Keep records, Help with decision making, Send alerts/warning. The farmer would: Map plots, Contact experts (including sending photos/videos), Receive warnings and Receive instructions about treatment applications. Concluding, the best opportunity for development would be a mobile application for Pests and Diseases Prevention and Control and the user profile that allows the creation of a more complex system would be the medium/large winegrower. Having said that, and as mentioned in 4.1.2, small farmers constitute the majority of producers in the Douro region, so it’s also important to take this into consideration and create an application that fits both user profiles. Figure F.1: Activity Setup 127 Brainstorming Activity 128 Appendix G Personas and Scenarios 129 José Pereira “Era caro e preferi perder algumas uvas do que estar a comprar os produtos. Ao preço que nos estão a pagar o vinho, se perder uma pipa ou duas compensa não o tratar” SCENARIO GOALS & TASKS age 75 years old role Smallholder winegrower education primary school mobile phone doesn’t own a mobile phone José learned everything he knows about his business from his father. He wakes up at 6AM everyday, he lives with his daughter and his grandson in a rural region of Portugal. His grandson helps him check if there are any alerts or releases for the region, provided by the SNAA (Serviço Nacional de Avisos Agrícolas) online, and he does so on his personal laptop. If needed he prints the information for José to read later. After arriving to his 1 hectare of land José starts by checking if there is any breakage on the vines, any visibto apply any products/pesticides he usually does so with the help from family members on the weekend and usually applies it by instinct and doesn’t follow the recommendations of the amount of treatment cycles due to the high costs. - Lose the least amount of vines - Track the maturation of the grapes to know the right time to harvest - Know how to prevent and treat pests and diseases (what products to apply) Personas and Scenarios 130 “Para competir com o mercado lá fora é preciso adotar as novas tecnologias, que permitem fazer coisas impensáveis há anos atrás. Quem não muda não evolui.” SCENARIO GOALS & TASKS Alfredo is one of the biggest wine producers in the Douro region, he inherited the business of his father and worked his way up and now owns over 70 hectares of land. He deals with a lot of pressure on a daily basis, his main motivation is to get a better understanding of the state of production and achieve the best possible quality wine. That being he procedes to send samples of grapes indicates the right time to begin harvesting each plot. After sending the samples to the lab, Alfredo waits for the results in his Email box or calls the lab if from the Summary Analysis, these are the most valued results, and then he receives the Supplementary Analysis. Also relating to the grape’s quality José relies on the same association that does the lab tests to get advice and visits relating to pests and diseases diagnosis and treatment. - Estimate production - Track the maturation of the grapes to know the right time to harvest - Detect signs that indicate pests or diseases (bugs, coloration of leaves) - Apply products/pesticides in localised applications - Achieve the highest quality of grape possible Alfredo Teixeira age 55 years old role Winegrower education higher education mobile phone owns an android smartphone Personas and Scenarios 131 Personas and Scenarios 132 Appendix H High-Fidelity Prototype Interface 133