Enhancing travel recommendations: Ai-driven personalization through user digital footprints

MBDESIGN - DESIGN, INNOVATION & TECHNOLOGY - UPC Romina Frias Velasco 2022 - 2023 “Enhancing Travel Recommendations: 1. A Digital World 1.1. The Digital Age 1.2. Digital Culture = Digital Society 1.3. Digital Identity 1.4. Digitalization of Services 2. Digital Services and Tourism 2.1. Modern day traveling 2.2. Tourism 4.0 2.3. Emerging Tourism Trends 3. High Tech  3.1.ArtificialIntelligence 3.1.1. Machine Learning, Neural Networks & Deep Learning 3.1.2. When Machines Speak Human 3.1.3. Garbage in. garbage Out 3.1.4. Aprehension & Enthusiam 4. Context of the Project 4.1. Online Presence and Privacy 4.1.1. Digital Footprint 4.1.2. Privacy Paradox 4.2. Motivations for Travel 4.2.1. Hierarchy of Needs 4.2.2. Push & Pull Factors 4.2.3. Travel Career Theory 4.2.4. Means-End Theory 4.3. Travel Decision Making 4.3.1. Social Proof 4.3.2. User-Generated Content 4.3.3. Electronic Word-of-Mouth 4.4. The Role of Personalization 4.4.1. The Paradox of Choice  4.4.2.TheCocktailPartyEffect 4.4.3. Personalization as a Solution 5. Case studies 5.1. Travel GPT 5.2. iplan.Ai 5.3. TripNotes 5.4. The Trip Boutique 5.5. Review of the examples 6. Inspiration 7. First steps 7.1. Empathy  7.2.Definition 8. The proposal 8.1. Ideation 8.2. Prototyping 8.3 Testing CONCLUSIONS 9. Final comments 10. Bibliography 11. Annex IMPLEMENTATION INDEX Introduction Research Questions Hypothesis Methodology CONTEXTUAL FRAMEWORK CONCEPTUAL APPROACH TECHNOLOGICAL FRAMEWORK 45 Acknowledgements I will be forever grateful to all those who have been a part of this journey. My professors, especially Enric and Josep María, whose invaluable guidance, expertise, and continuous support have shaped the work I’ve done throught this year My fellow classmates, with whom sharing the past months has been an immense pleasure. My family, partner and friends whose unconditional love, encouragement, and patience have pushed me through this year abroad. Thank you for beliving in me. Tutored by · Enric Trullols Farreny Typography · Freight Big Pro & Proxima Nova LT MBDesign · Design, Innovation & Technology Year 2022 · 2023 67 Abstract Abstract English Keywords Palabras clave Spanish This thesis aims to examine the way the digital footprint users leave behind can be utilized to optimize the personalization of tourism services, through the use of artificial intelligence. The paper proposes that the surge of artificial intelligence has opened a world of opportunities to develop new tools to improve the digital travel experience. The approach is based on the idea that digital footprints are unique and particular to each individual and this valuable data can result in smarter and unerring travel suggestions. Behavioral attitudes of the user, such as the influence of user-generated content in social media and e-word of mouth in the Esta tesis tiene como objetivo examinar la manera en que la huella digital que dejan los usuarios en internet puede utilizarse para optimizar la personalización de los servicios turísticos, mediante el uso de inteligencia artificial. El documento propone que el auge de la inteligencia artificial ha abierto un mundo de oportunidades para desarrollar nuevas herramientas para mejorar la experiencia de viaje digital. El enfoque se basa en la idea de que las huellas digitales son únicas y particulares de cada individuo y estos valiosos datos pueden dar lugar a sugerencias de viaje más inteligentes y certeras. Se consideran las actitudes de comportamiento del usuario, como la influencia del contenido generado por el usuario en las redes sociales y el boca a boca electrónico en el proceso de planificación del viaje, así como las implicaciones de este rastro de datos en la optimización de los servicios de viaje personalizados. Este modelo describe la relación entre la inteligencia artificial y la hiper personalización de servicios. Como es una tendencia creciente que está alterando nuestra realidad actual, la tesis presentada desarrolla una aplicación de viajes a medida que, con el permiso del usuario, aprovecha los datos recopilados de las redes sociales personales para construir un plan de viaje específico basado en las preferencias individuales. travel planning process, are considered, as well as the implications of this data trail in the optimization of customized travel services. This model describes the relationship between artificial intelligence and hyper-personalization of services. As it is a growing trend that is disrupting our current reality, the presented thesis develops a tailor-made traveling application that, with permission of the user, leverages the data collected from personal social media to build a specific travel plan based on each user’s preferences. Digitalfootprint,tourism,choiceparadox,artificialintelligence,machinelearning. Huelladigital,turismo,paradojadeelección,inteligenciaartificial,machinelearning. 89 Introduction Research Questions With the growing popularity of artificial intelligent chatbots such as ChatGPT, one cannot help but wonder about the great potential they bring. Between hopes and fears, a new reality is coming in which automation and digitalization are the new normal. We now are more aware of the use of artificial intelligence (AI), as it is no longer behind the scenes for product recommendations, not as our home assistant to play songs on command, not as bots that respond automatically. We now can use it to create songs, to write essays, to plan workouts and meals, to sort through pictures, to save ourselves precious time. As AI has never been as relevant as it is in this point is history, so is tourism after the global pandemic. People are still revenge-traveling, making up for the time missed in confinement and daring to visit locations never thought of before, especially now that restrictions are fairly relaxed in most countries. We now look for references through our social media, check reviews before purchase or reservation, trust e-word of mouth and have become content creators ourselves when we share our vacation photos and experiences with our online friends. Over the last two years, the travel industry has seen high demand of customers for digitalization, as these online services are perceived to make it easier for travelers to access information, book trips, and explore destinations. As the purchasing power of generations Y and Z, considered “digital natives”, is increasing, the greater the pressure for the digitalization of services. Today, digitalization is a must for companies that want to stay relevant in the current market, and with the growing trend of AI use, it is only a matter of time before we see hyper-personalization of services that perfectly suit each costumers’ individual demands. This thesis explores the importance of our individual digital footprint in the digital society we are living in, the way our interactions with people and things online influence our behavior and purchasing habits and the role of AI in personalization by analyzing the sea of data we create every day. Would people trust a tool that has personal data especially about behavior and preferences, by combining AI with social media, if they perceive the value obtained is worth it? Can we build a platform that considers multiple factors and variables for custom travel? Can travel suggestions come from people who have experience instead of companies trying to sell products, can they be based on each individuals’ interests instead of intentional mass-personalization marketing? 10 11 Hypothesis Methodology Artificial intelligence (AI) has become increasingly available for daily use over the last few years, in the form of personal assistants, movie platforms’ personalized recommendations to self-driving cars and advanced medical diagnosis tools. AI is changing our everyday lives by managing the data at scale and analyzing it to identify the newest trends, user behavior and preferences, to assist individuals and provide personalized recommendations. But the focus until not long ago was mostly marketing purposes for increased revenue. With the surge of AI-powered tools such as ChatGPT in the last few months, it is inevitable to wonder about the great potential of this technology for personal organization and leisure purposes. For this project, we will explore one of the possible future uses of AI, in a user-centered approach, by merging the potentiality of AI with the personal data footprint about personal interests gathered from social media. The hypothesis posits that through the intelligent analysis of individuals’ digital footprints, extraction of insights from user-generated content, and filtering of relevant information, both the integration For the development of this thesis, the design thinking methodology has been chosen. Interaction Design Institute and IDEO define it as a “problem-solving methodology that offers an approach centered around finding solutions.” it is a human-centered approach to innovation—anchored in understanding customer’s needs, rapid prototyping, and generating creative ideas. Design thinking is a systemic, intuitive, customer-focused problem-solving approach (McKinsey & Company, 2023), this methodology was chosen because it is seen as ideal for the nature of this project and because it places the needs and preferences of users at the center of the design process, ensuring that the resulting solutions are tailored to their expectations and requirements. The design methodology will follow a structured approach, incorporating the following key phases: empathy, define, ideate, prototype, and test. Each phase contributes to the iterative and user-centered nature of the design process. Empathy: During the empathy phase, user research will be conducted to gain deep insights into the preferences, pain points, and motivations of travelers. This will involve methods such as interviews, empathy maps, user journey, allowing for a comprehensive understanding of user needs and aspirations in the context of personalized travel experiences. of artificial intelligence (AI) with personal data footprints has the potential to revolutionize travel services by offering hyper-personalized travel suggestions. By reducing decision-making complexity, alleviating the paradox of choice, and providing efficient and enjoyable travel planning, it could be possible to enhance user satisfaction, foster greater engagement, and improve efficiency in travel decision-making processes. The exploration of theoretical foundations, relevant studies, and examples from the travel industry will shed light on the benefits and challenges associated with integrating AI and hyper-personalization in travel services. This user-centered approach, combining AI’s ability to process data at scale and understand user needs with the wealth of personal interests gathered from social media, could potentially create a seamless and tailored user experience. Define: The define phase aims to distill and synthesize the findings from the empathy phase to identify key design requirements and problem statements. This phase clarifies the design objectives, goals, and constraints, providing a solid foundation for the subsequent design ideation. Ideate: This phase involves brainstorming and generating a diverse range of design concepts and ideas. Prototype: The prototyping phase focuses on transforming selected design concepts into a tangible representation. This involves the creation of low-fidelity prototypes, such as sketches or wireframes, to visually communicate design ideas. Iterative prototyping will enable quick iterations and refinements, incorporating user feedback and ensuring alignment with the identified design requirements. Finally, a high-fidelity prototype, will be developed to simulate the personalized travel experience. Test: The testing phase involves gathering user feedback on the developed prototypes to evaluate their usability, effectiveness, and overall user experience. This user feedback is instrumental in identifying areas for improvement and refining the design solutions. In the context of this project, time represents a restraint that won’t allow for testing; however, it is important to highlight that nowadays, digital products are in a perpetual beta state, constantly testing and elevating the user experience of their services. 12 13 THEORETICAL FRAMEWORK 14 15 1. A Digital World The Digital Age, otherwise known as the Information Age or the Computer Age, is the present era where everything is dependent on the “widespread use of the Internet”. These advancements have transformed the way we live and work, and they have become a part of our everyday lives. The widespread use of the Internet is fueling one of the most exciting social, cultural, and political disruptions in history, a simple way to illustrate this is that as of June 2022, more than 500 hours of video were uploaded to YouTube every minute (Statista, 2023). This equates to approximately 30,000 hours of newly uploaded content per hour, with these numbers it is safe to assume that the amount of information available to us in this era is unprecedented. In order to understand the gradual growth of digital technologies, we can divide the Digital Age in three main periods as explained by Vorobiova, 2022: Pre-Digital: Refers to a time when technologies only served one purpose. Media such as television, newspapers, radio, and magazines were all one-way communication outlets. During this time, shopping was still the usual way to purchase products and services. Vorobiova illustrates these times using the example of physical phone books, that by now seem like a thing of the past as now that information is all stored on mobile devices. Mid-Digital: This is the phase we are currently in. As businesses are shifting towards digitalization of services, there is still a part of the population that remains trusting traditional technologies. A simple example is the preference for cable TV instead of subscription-based streaming services. According to the author we are still “lacking continuity”, in the sense that there is still a long way to go in the path towards digitalization. Post-Digital: She speculates that in this age the digital aspect will become hardly noticeable, just like electricity is for us at the present time. The term smart will be applied to pretty much every object we interact with, digital will just be a “fact of life”. This age will come with new challenges and opportunities. In the world of smartphones, people now have access to the world’s information at the tip of their fingers. In a way, it’s like they have “insurance against forgetfulness” (Schmidt & Cohen, 2013). “The Internet is the largest experiment involving anarchy in history. Hundreds of millions of people are, each minute, creating and consuming an untold amount of digital content in an online world that is not truly bound by terrestrial laws. This new capacity for free expression and free movement of information has generated the rich virtual landscape we know today.” Eric Schmidt & Jared Cohen, The new digital age Fig 1. Social Cut, Unsplash 1.1 The Digital Age 16 17 1.2 Digital Culture = Digital Society In the ever-evolving Digital Era, technological advancements have reshaped the way we live, communicate, and engage with the world around us. Central to this transformation is the emergence of a vibrant Digital Culture, which has fundamentally altered the dynamics of human interaction, self-expression, and knowledge sharing. As mentioned previously, culture and communication are two closely related concepts. Foresta et al., 1995 put special emphasis on two definitions from the Webster dictionary. The first is that culture is “the integrated pattern of human knowledge, belief, and behavior that depends upon man’s capacity for learning and transmitting knowledge to succeeding generations.” (Merriam-Webster’s Collegiate Dictionary, n.d.) The second concept is that culture refers to “the customary beliefs, social forms, and material traits of a racial, religious, or social group. “The characteristic features of everyday existence (such as diversions or a way of life) shared by people in a place or time”. (Merriam-Webster’s Collegiate Dictionary, n.d.) For Foresta et al., 1995 the difference between these two concepts is that the first one refers to the transmission of knowledge, while the second relates to the common rules for a community, their behavior, and the relationships they share with one another. (Foresta et al, 1995, as cited in Uzelac & Cvjetièanin. 2008). Building upon the foundation of Digital Culture, the Digital Society represents the broader social, cultural, and economic implications of digital technologies in our lives. It encompasses the transformation of various social structures, institutions, and practices due to the widespread adoption and integration of digital technologies. The Digital Society is characterized by the extensive use of digital devices, online platforms, and networks that facilitate communication, collaboration, and information sharing. Floridi, 2015 (as cited by Levin et al., 2021) elaborates on the three important “transformations” of the digital society. First is the blurred distinction between real and virtual, with this he implies that web presence has altered people’s notion of reality and it is increasingly difficult to distinguish the difference. The second is the blurring distinctions between nature, human and artifacts, this relates to the extensive integration of internet of things (IoT) and artificial intelligence in the human domain, almost considered as a “cognification” to our environment. And finally, the changing direction between scarcity of information to data overload, in a world where knowledge is abundant and easily available, what makes us survive in the sea of information is our capacity to pay attention and discern the false from the fact. Currently Digital Technologies surround us, they are part of the environment we live in, shaping the society we are part of, as we participate in virtual interactions. “Digitalization has enabled the process of media convergence to take place” (Uzelac & Cvjetièanin. 2008). These technologies foster an environment of interactivity and participation, where users are not only consumers of information but also creators of content. Today most of us find ourselves navigating through two realities simultaneously, the virtual one and the physical one, and we experience human reality through our mobile devices as most of our human interactions are done through that medium. They imply that culture is a “collective memory, dependent on communication for its creation, extension, evolution and preservation.” We can derive from the statements above that culture is strongly related to communication, and as the ways we communicate have changed over the last decades, we can agree that culture itself has changed too. At this point in time, all technologies mediate human experience and have transformed our society. Digital Culture encompasses the practices, values, behaviors, and forms of expression that have emerged within the context of the Digital Era. It reflects the ways in which technology has influenced and shaped our cultural norms, social interactions, and modes of self-expression. The digital landscape is characterized by the proliferation of online communities, virtual identities, and new modes of communication. Social media platforms, blogs, forums, and other digital spaces have become avenues for individuals to share ideas, experiences, and creative works, fostering a sense of belonging and collective identity within digital communities. Digital Culture is a dynamic and ever-evolving ecosystem that shapes and is shaped by the users who actively participate in it. Fig 2. Creative Christians, Unsplash 2020 30 31 “Ever since the arrival of printing - thought to be the invention of the devil because it would put false opinions into people’s minds - people have been arguing that new technology would have disastrous consequences for language.” “By far, the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.” —David Crystal, British linguist, academic, and author. —Eliezer Yudkowsky American artificial intelligence researcher and writer Fig 7 .Andrea De Santis, Unsplash 2020 3. High Tech To define what Artificial Intelligence (AI) is, it is important to first comprehend what the concept of intelligence entails. Amongst some of the definitions by the Merriam-Webster Dictionary, intelligence refers to the “ability to learn or understand or to deal with new or trying situations”, “the skilled use of reason”, and “the ability to perform computer functions”. For Rudas & Fautor, 2008 (as cited by Bulchand-Gidumal, 2020) intelligence envelops three crucial abilities: understand the environment and phenomena, take advantage of past experiences, and combine the knowledge available to respond appropriately to new challenges. AI is thereby defined as “the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable” (McCarthy, John, 2004 as cited by IBM, n.d.) A fundamental element to the function of AI is data because it supplies the essential input for AI systems to enhance through learning, identifying, and comprehending patterns of behavior, and finally, generate valuable insights. (Bulchand-Gidumal, 2020) Big data in the tourism industry is typically derived from two main sources: the environment and the tourists themselves. The environment serves as a valuable resource for meteorological data, real-time information from sensors, Internet of Things (IoT) devices, and transactions, as well as events happening at the destination. On the other hand, tourists contribute data in various ways, both before, during, and after their trip. This includes online activities, offline activities, biometric and emotional data, wearables, and User-Generated Content (UGC). Effectively utilizing UGC, requires processing with the help of Artificial Intelligence (AI) techniques. Sentiment analysis can be performed on textual information, enabling researchers and analysts to understand the sentiments expressed. Additionally, AI can analyze, and tag characteristics of pictures, audios, and videos shared by users, such as location, participants, and sentiments expressed. These AI techniques greatly enhance the value of UGC as a data source by providing more comprehensive information for data-driven processes. 3.1 Artificial Intelligence 32 33 In the scope of AI, there are two important and often mistaken with one another: Machine Learning (ML) and Deep Learning (DL) and in between the two, lies Neural Networks (NL). To better understand the framework in which this project will be developed, it is crucial to understand the difference between the three. Is considered a subdivision of artificial intelligence (AI) and computer science that mimics the way humans learn by using data and algorithms to be more accurate over time. Through the application of statistical methods and algorithms, ML classifies large amounts of data and makes predictions or discovers insights on certain projects. Some basic examples of everyday use of this tool are Netflix’s movie recommendations or self-driving vehicles. (IBM, n.d.) Classical ML is dependent on human intervention to learn, data scientists first need to determine and structure the data for it to be correctly processed. Depending on the degree of human intervention on raw data, there are four types of ML (UC Berkeley, 2020): UC Berkeley also develops on three main components to the process of ML: - A decision process: The algorithm makes estimates about patterns from multiple guesses and calculations. - An error function: Assesses the accuracy of the model by comparing it to known examples. - An updating or optimization process: If the model can accurately fit the data it’s given, it adjusts its “settings” (weights) to make its predictions match the actual examples more closely. The model keeps going through this “evaluate and improve” process, tweaking its settings on its own until it gets as accurate as possible based on a specific target. There are multiple ML algorithms used for a wide variety of purposes, amongst which the most used are: Linear Regression, Logistic Regression, Decision Trees, Random Forests and, the one pertinent to the development of this project, Neural Networks. Neural networks are AI algorithms that understand and classify data in a smart manner by mimicking the thought process of a human brain, by recognizing patterns in data and speech through massive amounts of intertwined processing nodes (UC Berkeley, 2020). Data is processed through layers, with each layer assigning weights before passing it to the next layer. This way the algorithm quickly learns and adjusts itself for more efficiency. Developing more on this process IBM, n.d. defines NL basic structure as composition of three node layers: first an input layer, one hidden layer and lastly an output layer. These nodes work as artificial neurons that are connected to each other and have weights and thresholds. When the output of a node exceeds the threshold, it gets activated and passes data to the next layer. If the output is below the threshold, no data is sent to the next layer by that node. This way, information flows through the network based on whether nodes are activated or not. If the number of layers in NN consists of more than three, including the input and output, then it is considered a Deep Learning algorithm. - Supervised learning: The dataset is labeled and classified before use to allow the algorithm to see how accurate its performance is. - Unsupervised learning: The algorithm pinpoints patterns and similarities from unlabeled data. - Semi-supervised learning: The algorithm makes independent conclusions as it is fed both structured and unstructured data, this way the algorithm learns to label data itself. - Reinforcement learning: The algorithm learns from experiences by trial and error and gets feedback in the form of rewards and punishments. 3.1.1 Machine Learning, Neural Networks, Deep Learning Machine Learning Neural Networks Figure 8. Kavlakoglu, Eva (May, 2020) “AIvs.MachineLearningvs.DeepLearningvs.NeuralNetworks:What’stheDifference?” Figure 9. Nielsen, Michael (December 2019) “WhyareNeuralNetworkshardtotrain?” 34 35 The term “deep” in the context of Deep Learning pertains to the depth of layers within a neural network. Specifically, if a neural network encompasses more than one hidden layer (UC Berkeley, 2020), it is considered a Deep Learning algorithm. The depth of the network is indicative of its ability to process and analyze data through multiple layers of abstraction, enabling more sophisticated and intricate representations of information. Consequently, the inclusion of multiple hidden layers distinguishes Deep Learning as a distinct paradigm within the field of machine learning. With DL the process of extracting important features from the data is automated, reducing the need for manual human involvement. Currently it also referred as “scalable machine learning.” because it works perfectly with large datasets. A simple example of this is illustrated by MIT, 2021: “in an image recognition system, some layers of the neural network might detect individual features of a face, like eyes, nose, or mouth, while another layer would be able to tell whether those features appear in a way that indicates a face.” Just Like NN, Deep Learning is designed to mimic the thought process of the human brain and is widely employed in various applications of machine learning. This technology drives advancements in areas such as autonomous vehicles, chatbots, and medical diagnostics. Deep Learning (DL) 3.1.2. When Machines Speak Human As a subfield of AI, Natural Language Processing (NLP) is focused on providing computers with the capability to comprehend text and spoken language like humans. (IBM, n.d.). According to IBM, NLP “merges rule-based computational linguistics with statistical, machine learning, and deep learning models”. These technologies allow computers to process human language and understand complex concepts such as intent and sentiment, in the form of text and voice. In the average daily basis, NLP can be found in digital assistants, GPS systems or customer service chatbots. Since this technology addresses specifically human language, there is a series of tasks it need to perform in order to decipher the complexity of people’s expressions: metaphors, sarcasm, homonyms, homophones, to name a few. Citing IBM’s definition of NLP, among these tasks are: - Speech recognition, also known as speech-to-text, involves converting voice data into text data and is crucial for applications that rely on voice commands or respond to spoken questions. - Grammatical tagging, that determines context and use in speech. - Word sense disambiguation, which refers to the selection of the correct meaning of a word, depending on the context. - Name entity recognition identifies important entities like locations and names. - Co-reference resolution identifies when two words refer to the same thing or entity, or interpreting a metaphor, slang. - Sentiment analysis, which extracts subjective qualities such as emotions and sentiments. -Natural Language Generation, that puts structured information in comprehensive human language. Some uses of NLP include virtual agents and chatbots which recognize prompts and commands from human requests and in return provide relevant answers; another tool is Social Media Sentiment Analysis, which has become a crucial tool for businesses to gain insights from social media data. With this method. attitudes and emotions towards certain products/services can be extracted from social media posts and reviews. Another use worth mentioning for NLP is text summarization, this tool sorts through vast amounts of text and provides summarized synopsis of databases. Figure 10. Karnes, KC (2019) “Introduction to (NLP) 36 37 Figure 11. Baheti. 2021. “A Simple Guide to Data PreprocessinginMachineLearning” Figure 12. Cser, Tamas (2018) IsAIGoodforSociety?-TheGood,TheBad&TheUgly 3.1.3. Garbage In, Garbage Out 3.1.4. Apprehension & Enthusiasm In the realm of data analysis and machine learning, the phrase ‘garbage in, garbage out’ has become a well-known mantra. It encapsulates the critical importance of data quality in determining the reliability and effectiveness of any analytical process or model. It is a common understanding in ML that the better the amount of data, the better the models are trained. Unfortunately, real-world data is often polluted with imperfections such as inconsistencies, noise, incomplete information, and missing values. It is derived from diverse sources through the utilization of data mining and warehousing techniques. (Baheti,2021). For this reason, data cleaning is mandatory for a successful result. Data pre-processing, as a fundamental step in the data pipeline, plays a pivotal role in transforming raw and unrefined data into valuable insights. By meticulously cleansing, transforming, and enhancing the data, pre-processing ensures that only high-quality, reliable, and relevant information is fed into subsequent analytical tasks. There are 4 main techniques used for data pre-processing (Baheti, 2021): Data Cleaning: Involves identifying and handling any errors, inconsistencies, or missing values in the dataset. It may include techniques such as imputation, removing duplicates, and correcting data errors. Data Integration: In this step, data from different sources or formats is combined into a single cohesive dataset. It may involve resolving naming conflicts, data format standardization, and merging relevant data. Data Transformation: This step involves converting the data into a suitable format for analysis. It may include scaling, normalization, encoding categorical variables, and feature engineering to create new informative features. Data Reduction: This step focuses on reducing the dimensionality of the dataset by selecting relevant features or applying techniques such as dimensionality reduction or feature selection. This helps in reducing computational complexity and noise in the data. The rise of generative AI models like ChatGPT, DALL-E, and MidJourney has sparked both fascination and apprehension. These models showcase impressive capabilities in generating human-like text, images, and music, but also raise concerns about systemic bias, privacy, misinformation, malicious use, and displacement of human labor. The development of regulatory frameworks and ethical guidelines lags behind the rapid advancement of AI, calling for responsible and transparent use. To strike a balance, we should embrace the potential benefits of generative AI while addressing legitimate concerns. Open discussions, interdisciplinary collaboration, and appropriate regulations can help mitigate apprehensions and ensure ethical use. AI can increase efficiency, bridge gaps, improve accessibility, revolutionize healthcare, fuel innovation, and enhance user experiences. It is crucial to remember that AI is a tool developed and used by humans. Responsible development, ethical considerations, and ongoing research are necessary to align AI with human values. By being educated, fostering critical thinking, and advocating for ethical practices, we can harness the power of AI to drive innovation and improve lives. 38 39 CONCEPTUAL APPROACH 40 41 “You can’t talk about big data without talking about things like privacy and ownership.” —Rick Smolan, photographer, and co-author of The Human Face of Big Data Fig 13 . EV, Unsplash 2018 4. Context of the Project In today’s digital age, our lives have become intricately intertwined with the online world. We share our thoughts, experiences, and personal information on various online platforms, connecting with friends, family, and even strangers across the globe. However, this increasing online presence raises important concerns about privacy and the protection of our personal data. As we navigate the vast landscape of social media, online shopping, and digital communication, we leave behind a trail of digital footprints—traces of our activities, interactions, and preferences. These footprints, though seemingly innocuous, can paint a detailed picture of our lives, interests, and behaviors. Our digital footprints are compiled, analyzed, and often monetized by companies and organizations, shaping the online experiences we encounter. This interplay between our digital presence and the protection of our privacy gives rise to what is known as the privacy paradox. On one hand, we crave the convenience, connectivity, and personalized experiences offered by technology. On the other hand, we value our privacy and the control over our personal information. Navigating this paradox becomes a delicate balancing act as we weigh the benefits of sharing information against the potential risks of misuse and intrusion. 4.1. Online Presence and Privacy 42 43 4.1.1. Digital Footprint 4.1.2 Privacy Paradox As stated before, individuals currently possess two identities, a physical identity that can be verified by official paper documentation, and digital one that is refers to an individual’s interactions online, including search history and social media (Park, 2017 as cited by Blue et al 2018). These online interactions leave behind what is called a digital footprint. A digital footprint indicates a person’s online existence and serves as proof of their digital and physical identities. It records the trail and artifacts left behind by people engaging in a digital environment (Fish, 2009 as cited by Blue et al, 2018). As digital footprints map and record more parts of an individual’s real-world existence, they provide valuable information and qualities about subjects such as professional affiliations, social relationships, personal health information, purchases, habits, interests, and much more. (Blue et al, 2018). In the case of travel digital footprint, social media posts of food, landmark visits, leisure activities, check-ins into hotels or cities, pages followed, interactions made with content such as videos watched and liked regarding travel and tourism, make for meaningful data that can reveal a lot about what kind of travel each individual is interested in. Whenever individuals use online services, they make a decision to give up their privacy in return for benefits. People, for example, are sometimes required to identify their location to obtain a real-time weather prediction or to share their preferences to get appropriate suggestions for products or activities. Indeed, most free mobile services’ economic models are based on similar trade-offs: data is the currency in which users pay for online services. (Liu & Simpson,2020) Privacy concerns are an essential part of the digital identity formation process, especially because these concerns may prohibit information disclosure. (Papaioannou, et al 2020). Understandably, users have concerns about how much data they are really disclosing and what happens to this data once it is property of a third party. However, some research suggests that online users’ privacy concerns may vary depending on the situation, aspects such as perceived benefits, financial cost, privacy risk, take part in users’ final decisions of information disclosure (Liu & Simpson,2020). When users are faced with the dilemma of whether the perceived benefits of using online services are worth trading personal information or not, their behavior can vary substantially from their self-reports. Users who claim to have serious concerns about privacy risks might also voluntarily share their personal information for certain rewards or allowances. (Liu & Simpson,2020). Liu and Simpson assert that there are three main factors that directly influence the willingness of users to exchange personal information for online services. Before and during their trip, tourists engage in online searches and bookings for services, leaving a digital footprint that can be tracked. Additionally, they leave offline traces such as movements, booking records, and consumption patterns, which can be captured through GPS data, mobile roaming, Bluetooth devices like beacons, IoT devices, and Point of Sale (PoS) systems. Biometric and emotional data, such as thermal images and facial recognition, can also be automatically collected during the trip. Moreover, depending on the user’s consent, data from wearables like smartwatches, activity trackers, and clothing can be gathered. Lastly, UGC is generated during and after the trip, including online reviews, social media comments, and shared pictures and videos. (Bulchand-Gidumal, 2020) User profiles result from the combination and analysis of this data and is used to offer personalized recommendations for products and services that cate each users’ needs. (Bulchand-Gidumal, 2020) First, “awareness and knowledge of privacy risks”, clear perception of privacy practices in digital services allows users to make mindful decisions about data-sharing, However, their research suggests that users who are unaware of privacy protocols are more likely to share personal information out of ignorance. To add on, findings point to the fact that some users who regard themselves as wary of privacy policies, would unconsciously ignore these beliefs when faced with a worthy trade-off. Second is “trust in service providers” meaning the more reputable the enterprise, the more likely users would feel comfortable sharing information with it. Research findings hint that trust in the prestige of a company “weakens users’ awareness of privacy risks” (Liu & Simpson,2020). Lastly, “desire for mobile services”, aiming that a strong necessity for special and unique digital services may incline users to overlook their privacy concerns in favor of getting their needs fulfilled. (Liu & Simpson,2020). Another interesting factor worth mentioning is what Liu & Simpson call “beliefs of cyber privacy”, which touches on the users’ assumption that cyber security does not exist, they have got nothing to hide or even if their information is shared, they don’t regard their data as “valuable”. We can conclude that privacy is a decisive aspect of digital tool use, however, while users appreciate transparency and honesty, they are willing to trade data in certain situations when the digital tool is perceived to be beneficial to them. 44 45 Human beings have an inherent desire to explore and discover the world around them. From ancient times to the modern era, travel has played a significant role in fulfilling this fundamental human need. The motivations that drive individuals to embark on journeys are diverse and multifaceted, ranging from basic physiological needs to more complex psychological and emotional aspirations. Understanding these motivations is crucial in comprehending why people choose to travel, how they make their travel decisions, and what experiences they seek Several theories and frameworks have been developed to shed light on the underlying factors that shape our travel motivations. We will explore prominent theories to gain valuable insights into the psychological, social, and cultural forces that influence our travel choices. They help us recognize that travel is not merely about reaching a destination but a journey of self-discovery, personal growth, and fulfillment. By delving into these theories, we can gain a deeper understanding of the complex interplay between individual desires, external influences, and the quest for meaningful experiences. 4.2. Motivations for Travel “I have wandered all my life, and I have also traveled; the difference between the two being this, that we wander for distraction, but we travel for fulfillment.” “Of the gladdest moments in human life, methinks, is the departure upon a distant journey into unknown lands. Shaking off with one mighty effort the fetters of Habit, the leaden weight of Routine, the cloak of many Cares and the slavery of Civilization, man feels once happier.” — Hilaire Belloc, poet, historian, essayist – Richard Francis Burton, explorer, writer, orientalist 4.2.1. Push & Pull Factors When discussing motivations for travel, it’s essential to consider the factors that push individuals to leave their current location and those that pull them towards new destinations. These factors, commonly known as push and pull factors, help us understand the underlying reasons for people’s travel choices. Push factors encompass a range of circumstances, motivations, and stimuli that exert a “pushing” effect, prompting individuals to seek alternatives to their current location. These factors typically arise from a sense of discontent with the prevailing conditions, whether it be dissatisfaction with the local environment, economic constraints, political instability, safety concerns, or a general desire for change. Push factors act as catalysts, generating a compelling impetus for individuals to explore new horizons, escape the limitations of their current situation, or pursue personal growth and enrichment through travel. Push factors are the underlying forces that direct an individual’s decision to travel (Chen & Chen, 2015 as cited by Jumrin & Maryono, 2018), they encompass psychological motivators such as social interaction, the yearning for escape, adventure, relaxation, and self-exploration, and many more (Annex 3) In contrast, pull factors represent the alluring forces that draw individuals towards specific destinations, captivating their attention and sparking a desire to visit. These factors encompass the attractions, opportunities, and positive aspects that a particular location offers. This refers to the destination image and it is represented by the complete perception of the visitor towards the tourist site. (Lee, 2009 as cited by Jumrin & Maryono, 2018) . Perception holds a significant influence and acts as a pull factor that drives tourists to select certain destinations, and according to Cherry, 2003 (as cited by Jumrin & Maryono, 2018) there is an intricate link between perception and the stimulus received from the surrounding environment, subsequently encouraging individuals to react in the form of activities. Cultural richness, historical significance, natural beauty, recreational pursuits, favorable climatic conditions, renowned landmarks, and the reputation for warm hospitality are among the many elements that contribute to the magnetic pull of a destination. Pull factors play a pivotal role in inspiring and captivating the imaginations of individuals, enticing them with the promise of novel experiences, adventures, and opportunities for self-discovery. (Annex 3) The relationship between decision-making and push and pull factors is a nuanced and intricate one. As individuals engage in the process of travel planning, they carefully evaluate and weigh the influence of these factors, considering the interplay between their personal circumstances, aspirations, and the allure of various destinations. Push factors serve to incite the need for change, motivating individuals to seek out alternative environments or experiences. Simultaneously, pull factors provide an attractive vision of what awaits at the chosen destination, encouraging individuals to pursue their travel aspirations with enthusiasm and anticipation. 46 47 4.2.2. Travel Career Theory 4.2.3. Means-End Chain The Travel Career theory, developed by Pearce, 1991 (as cited by Jiang 2019), is a hierarchical system consisting of five levels, designed to rank and categorize the post-travel motivational descriptions provided by tourists regarding their holiday experiences. It was inspired by Maslow’s theory and at its core, is the idea that people’s motivations and preferences for travel change as they gain more experience and knowledge about different destinations and travel styles. The Travel Career Ladder (TCL) recognizes that travelers move through different stages or levels, each characterized by distinct motivations and behaviors. The model proposes that people have a specific career goal in their tourism behavior and as they gain more experience, their focus shifts towards fulfilling higher-level needs and seeking greater satisfaction. (Pierce 1991, as cited in Ryan, 1998) The steps of the ladder (from bottom to top) are physiological needs, safety/security, relationship needs, self-esteem/development needs, and fulfillment needs. Later, Pearce (2011) improved his theory by developing the Travel Career Pattern (TCP), see Annex 4. The Means-End Chain provides a framework for understanding the relationships between personal values, product attributes and consumer benefits or consequences, and. (Gutman, 1982 as cited by Jiang 2019). Personal values are the underlying principles or beliefs that individuals hold dear. These values guide and shape their preferences, behaviors, and decision-making processes. In the context of the MEC theory, personal values are the ultimate ends or goals that individuals seek to fulfill through the consumption of products or services. According to Rokeach, 1973 (as cited by Jiang 2019), there are two types of personal values: -Instrumental Values: Which are about the way we behave or act. They describe the qualities or behaviors that we consider important or desirable in ourselves and others. These values are like guidelines for how we should ideally conduct ourselves. For example, instrumental values could include honesty, kindness, hard work, or fairness, in the context of travel, they could include sustainability, reliability, honesty, etc. - Terminal Values: Are about the end results or outcomes that we aspire to achieve in life. They reflect the ultimate goals or states of being that we consider valuable and meaningful. Terminal values are related to our desired end-states of existence. For example, happiness, success, inner peace, love, or personal fulfillment, and in the context of travel, they could range from relaxation, exploration, connection to personal growth. Next, Attributes refer to the perceived physical or abstract characteristics of a product/service (Gutman, 1997, as cited by Jiang, 2019). In the tourism industry, Jiang recognizes two relevant types of attributes: - Concrete: An objective property of a destination (local customs) - Abstract: Relative, reflective attribute (fame, perceived environment) Finally, Consumer benefits are the positive outcomes or consequences that consumers associate with these attributes and result from their behaviors (Gutman 1982, as cited by Jing, 2019). Benefits are the reasons why consumers value certain attributes of a product. The MEC theory suggests that there is a hierarchical relationship between these three components, (Jiang, 2019) and it proposes that attributes are means to achieve certain benefits, and these benefits, in turn, serve as means to fulfill personal values. 48 49 In today’s digital age, the influence of social proof has become increasingly important in the travel industry and has a profound impact on how individuals perceive destinations and make decisions. It is crucial to recognize the way the landscape has evolved with the advent of digital platforms and social media, and how these technologies offer individuals valuable insights, authentic experiences, and peer recommendations that influence their perception of destinations and ultimately impact their choices. Rather than relying solely on traditional sources of information, such as travel agencies or official destination websites, travelers now turn to social media platforms, online reviews, and recommendations from fellow travelers to inform their decision-making. 4.3. Travel Decision Making “If you make customers unhappy in the physical world, they might each tell 6 friends. If you make customers unhappy on the Internet, they can each tell 6,000 friends.” – Jeff Bezos, CEO at Amazon.com 4.3.1. User Generated Content As communication channels have evolved through time, people search for inspiration and information in the available channels, with the rise of social media, paper magazines and travel encyclopedias are a thing of the past. Since social media allows users to share their experiences, it has become one of the most popular sources of information for travelers about their potential destinations, many of them being content created by travel Social Media Influencers” (SMI). Horton and Richard (1956) defined parasocial interaction as a face-to-face illusionary link between the audience and media icons such as celebrities, artists, and broadcasters, it is described as the establishment of a cordial relationship between customers and personas (characters or media figures) via the use of certain media. In the same way, SMI’s presence is more affordable if they create a positive bond with their followers. Apart from the importance of aesthetics, the authenticity of information shared will determine the strength and success of the parasocial interactions. Consequently, parasocial interactions have a direct influence on followers’ behavior and consumer decision. (Cheng et al, 2023) Fig 14 . Karsten Winegeart, Unsplash , 2021 62 63 “Information is a source of learning. But unless it is organized, processed, and available to the right people in a format for decision making, it is a burden, not a benefit.” —William Pollard, American physicist Fig 24. Tom Barret Unsplash 2017 6. Inspiration The rising popularity and advancements in AI technology have garnered significant attention across diverse industries. This technology possesses the capability to generate content, simulate human-like behavior, and deliver personalized experiences, thereby opening up new avenues for innovation. Concurrently, social media platforms have emerged as influential forces in shaping consumer behavior and purchase decisions. People actively share their preferences, interests, and experiences on these platforms, thereby generating a substantial amount of data that reflects their individual tastes. Through careful observation of my own social media presence, as well as that of my peers, it becomes evident that a tremendous volume of valuable information is produced on a daily basis. In today’s digital landscape we find ourselves continuously engaging with content. While some of this content is shared with friends, other pieces are deemed valuable and worthy of saving for future reference and go straight to the “save for later” album/file. Amidst the rapid flow of information, this later (most of the time) never comes. This once valuable information gets buried under the newer, more relevant next best thing. Fig 25. Instagram Saved Content , Travel Files. 64 65 There is a transitory nature to our digital realm, but despite of that we could acknowledge the inherent potential that resides within the copious amount of saved content and the digital footprint we leave behind. This repository of preserved information encapsulates a wealth of valuable insights and experiences, awaiting harnessing and organization in a manner that can genuinely benefit us. Traditionally, businesses have collected user data without always exhibiting full transparency regarding its utilization. However, what if we could establish a more direct and upfront data collection relationship lies in transparency? By clearly communicating to users that their social media data will be used for their direct benefit, individuals can make informed decisions regarding the sharing of their information. This approach could foster trust and empower users to actively participate in the data collection process, offering them the choice to share their data in exchange for the personalized services they receive, specifically tailored travel suggestions. Through conscientious curation and organization of our saved content, we are presented with an opportunity to construct a personalized knowledge base, serving as a powerful tool for personalization and enabling the leveraging of our own digital footprint to enhance our life experiences. In this proposal, where users collaborate with the final product to augment the value delivered, the co-creation of value becomes a crucial aspect. By contributing their personal data, users actively participate in shaping the personalized travel recommendations they receive. They become co-creators of value by offering insights into their preferences and contributing their data and trust. Advanced AI techniques then process their data to deliver highly personalized recommendations. Recognizing the influential capacity of social media, I have explored the concept of utilizing this data to enhance personalized travel recommendations. The objective is to develop a practical and efficient tool that furnishes customized travel suggestions based on individuals’ distinct interests and preferences. This amalgamation of technologies and data holds significant promise in creating valuable, user-centric solutions within the travel and tourism domain. Fig 26. Instagram liked Content , Travel Related Fig 27. Social Media, Sara Kurfeß, Unsplash 2018. 66 67 “Design thinking relies on our ability to be intuitive, to recognize patterns, to construct ideas that have emotional meaning as well as functionality, to express ourselves in media other than words or symbols.” —Tim Brown, Change by Design Fig 28. Jason Goodman Unsplash 2019 7. Design Thinking In the quest to create a personalized travel app that caters to the unique preferences and needs of individual users, it is imperative to analyze the target audience, to understand their thinking and behavior. This understanding forms the foundation of the empathy phase in our design thinking process. First, the target audience of this services where identified as generations Z and Millennials who are the first generations to grow up with technology at their fingertips, characterized as highly tech-savvy and comfortable using digital platforms and mobile apps. Amongst other peculiarities, they are heavily influenced by social media when it comes to travel decisions. They actively engage with social media platforms like Instagram, TikTok, and YouTube to seek inspiration, gather travel information, and share their experiences, and finally they value personalized experiences. A profile of both generations and their expectations from digital services is created to better understand their needs. 7.1. Empathy Fig29.GenerationYProfile,InspiringApps. 68 69 Interviews were conducted to understand travelers’ thought process during their travel planning phases, and the reasons why they chose their destination. The sample was meant to include people from several backgrounds and nationalities within the age group the service is aimed for, which is millennials and Gen Z. While there were inherent limitations of a sample predominantly composed of Bolivian participants, efforts have been made to justify and mitigate these biases. This includes the incorporation of additional data sources to understand the behavior of travelers from other nationalities. The data from those interviews and from the research on other countries traveler profiles (Annex 5), confirmed the previous research done on motivations to travel. Every person interviewed or traveler profile had very particular motives and travel planning processes, depending on their interests and travel career. For example, for those who lived in a country without ocean access, the most sought-after destinations were places with beaches and warm weather. Those who were interested in music put festivals and concerts above all else. While some would rather pay for a travel agency service to not deal with the hassle of planning, many just go to social media for reference and do the planning themselves. Most of the interviewed strongly rely on social proof for reference before purchasing or making a reservation, although some now mistrust social media due to the over-glamourized representation of travel. Some suggested the idea of “Smart Recommendations” that could arise after data input, such as high and low budget options, festivities and events that will be happening during the dates of the trip or recommendations that consider all constraints (budget, time, preferences) or simultaneously. Another interesting feature of value mentioned was the centralization of information, handy Qr codes for flights and reservations stored in one place, for easy access. User Interviews Fig30.GenerationZProfile,InspiringApps. 70 71 During this phase, all the information collected from the research and interviews in synthetized in a graphic manner to better understand user common characteristics, use patterns and challenges. A user persona is created as a fictional representation of the target audience, in this case millennials, where a personal story is included to bring the persona to life and illustrate their goals and challenges. User Persona helps us develop a deeper understanding of their characteristics, behaviors, needs, and motivations 4.3. Travel Decision Making “Your most unhappy customers are your greatest source of learning.” – Bill Gates, American business magnate, investor, and philanthropist. 7.2 Define User Persona Fig 32. Brooke Cagle Unsplash Fig 31. Interviews insight 72 73 Key Takeaways: To effectively cater to millennial travelers like Anna, travel service providers should focus on providing unique and authentic experiences, leveraging technology and digital platforms, offering budget-friendly options, promoting sustainable practices, and fostering a sense of community and trust through user reviews and recommendations. Fig 33. User Persona 74 75 Customer Journey Map The customer journey map focuses on mapping out the actual steps and touchpoints of the user’s interaction with the final value proposal, in a way acting as a bridge between the user understanding and user experience. This visual representation illustrates pain points, gaps, or opportunities for improvement within the user’s journey. By visualizing the entire end-to-end experience, areas where user experience falls short are identified. Some important points are apprehension about new technology, privacy concerns, price convenience concerns, personalization capacity. Empathy Map In the definition phase of design thinking, the empathy map serves as a valuable tool to understand and empathize with the users of a travel app. Its purpose is to gain deep insights into the thoughts, feelings, motivations, and behaviors of the app’s target audience. By creating an empathy map, we can step into the shoes of the users, gaining a clearer understanding of their needs, desires, and pain points. Important aspects identified are the role of social media, information overload, and choice paralysis. Fig 34. Empathy Map Fig 35. Customer Journey Map 76 77 “The best way to create value in the 21st Century is to connect Creativity with Technology” —Steve Jobs, business magnate, inventor, and investor. Fig 36. Jason Coudriet Unsplash, 2018 8. The proposal During the ideation phase of the project, I engaged in a creative and exploratory process to generate innovative ideas and possibilities for the travel app. One of the techniques I employed was the SCAMPER method. Using SCAMPER, I thoroughly evaluated each element of the existing travel planning journey and identified opportunities for improvement and disruption. I explored various ideas for substitution, such as replacing traditional booking methods with alternative options like peer-to-peer accommodations or blockchain-based systems. I also considered combining functionalities to create a more seamless and integrated experience, leveraging emerging technologies such as artificial intelligence. 8.1. Ideation 78 79 Fig 37. Scamper method for ideation Throughout the ideation process, I explored various iterations of the potential functions the travel app could offer. Then the questions that came up were: What exactly is the app? What role does it play in the travel journey of our users? Could it be more than just a tool? In a way it acts as an ally, an assistant, a digital travel companion that understands and caters to the unique needs and preferences of our users. Drawing inspiration from the metaphorical potential, the aim is to create experiences that go beyond traditional travel planning methods. Fig 38. Ideation mind map 80 81 • As an ally: the service could provide unwavering support throughout the entire travel journey. It offers guidance, recommendations, and solutions to the challenges faced by travelers. It becomes a reliable partner, ensuring that every step, from destination selection to itinerary planning, is seamless and enjoyable. • As an assistant: the service could take on the role of a helpful travel companion. It assists users with various tasks, such as booking accommodations, finding local attractions, and suggesting personalized itineraries. With intuitive interface and intelligent algorithms, it can simplify the complexities of travel planning, making it effortless and enjoyable. • As digital travel agent: It could leverage its vast knowledge base and user data to provide tailored recommendations and suggestions. By understanding users’ preferences, past travel experiences, and interests, it offers personalized suggestions for destinations, activities, and hidden gems, creating a truly customized travel experience. • As a friend: a trusted confidant who knows users intimately. It draws on its knowledge of users’ preferences, digital trail, previous trips, and social connections to offer recommendations that resonate on a personal level. It becomes a friend who shares their wisdom, insights, and experiences, creating a sense of familiarity and trust. During the prototyping phase of the project, my focus shifted towards exploring the most suitable AI techniques that could be applied to enhance the functionality and user experience of our travel app. Recognizing the potential of AI in revolutionizing the way travelers plan their trips, I embarked on a research journey to identify the AI techniques that align with our project goals. With an emphasis on personalization, recommendation systems, and intelligent data processing, I delved into various AI methodologies such as machine learning, natural language processing, and data analytics. • Data Collection • Data Preprocessing • NLP Techniques • Supervised Machine Learning • Feature Engineering • Model Training • Collaborative Filtering • Personalized Travel Recommendations. • User Interface and Interaction However, it is important to acknowledge that despite the comprehensive research and understanding gained during the research phase, the actual implementation and development of these AI techniques require a deep level of expertise in coding and database management. It becomes evident that the execution of the envisioned AI-powered features and functionalities will require the collaboration and involvement of skilled developers and data specialists. A detailed graphic about the steps to develop the project in terms of the basics is presented in a step-by-step manner for easier understanding about what every phase entails. And a rough algorithm is presented in a very simplified version of what a fully-fledged system would entail. However, it serves as a proof of concept and provides a starting point for further refinement and optimization. It allows us to understand the potential of using algorithms to enhance the travel planning process, guiding users towards personalized and memorable experiences. (Annex 6) 8.2 Prototyping “If a picture is worth a thousand words, a prototype is worth a thousand meetings” ― IDEO.org The technology behind 94 95 Final Proposal A mobile application that shows traveleres a tailored world fit to their liking and personality. The welcome screen presents the main idea of the service, while the other four coaching screens explain the basic steps they must follow in order to get curated travel suggestions. The screens above asking for two step verificiation to ensure account protection. Social pairing screen, where users get to choose which social media they want to pair with the app. Placing a high importance in privacy, they must first read and agree to the privacy details. Figures 40 through 43. Prototype 96 97 Personalization screens, that fit each different type of trip, amongst the filtes are insterests, budget, time available, type of acommodation and flight preference and companions. Later a loading screen while the app. elaborates the travel proposals. The home screen presents potential destinations that could be of interest for the user, based on their travel behavior, history and digital trail, where they can save for later or simply like. Once they choose, an elaborated proposal is developed and presented that they can also tailor to their liking as it automatically suggests replacements. 98 99 While the ideal design thinking process involves thorough testing and iteration of prototypes to gather user feedback and refine the product, time constraints and limited knowledge in data processing, coding, and the required back-end development and AI technology, have limited the scope of this project, preventing the inclusion of a dedicated testing phase. While the basic outline of the personalized travel app based on social media footprint has been established, the absence of the necessary technical expertise hinders the ability to fully implement and test the solution. The development of a robust back-end infrastructure, encompassing data collection, storage, analysis, and recommendation generation, requires specialized skills and knowledge in areas such as database management, algorithm development, and machine learning. process would allow for the identificati These technical components are essential for accurately processing social media footprints, extracting relevant information, and delivering personalized travel recommendations. In a typical testing phase, prototypes would be developed and presented to target users, who would provide valuable feedback through observations, interviews, and usability tests. This iterative process would allow for the identification of potential issues, user preferences, and areas for improvement, ultimately leading to a more refined and user-centric design. However, given the limitations, the focus of this thesis has been on the earlier stages of the design thinking process, such as research, empathy, and ideation. These stages have provided valuable insights into the user needs, motivations, and desired features, which have resulted in the design of a mobile application user interface. 8.3 Testing “I think test-driven design is great. But you can test all you want and if you don’t know how to approach the problem, you’re not going to get a solution.” ― Peter Norvig, computer scientist and education fellow at Stanford Institute for Human-Centered AI. Fig 44. Faizur Rehman Unsplash, 2021 100 101 CONCLUSIONS 102 103 “Though we see the same world, we see it through different eyes.” —Virginia Woolf, writer. Fig. 45 Clay Banks Unsplash, 2017 9. Final Comments In addition to its direct implications for personalization of travel experiences, this project could open a broader discussion on the potential for data cross-collaboration and co-creation of value in various fields. The concept of personalization, when implemented with explicit and transparent consent, has the potential to revolutionize industries beyond travel. By leveraging social media footprints and user-generated data, organizations can gain valuable insights into individual preferences, behaviors, and aspirations. This data-driven approach allows for the creation of tailored experiences, products, and services that meet the unique needs of everyone. Involving users throughout the design process of their experiences, and incorporating their insights fosters a sense of ownership and ensures that the final product aligns with their expectations and aspirations. This co-creation approach empowers users, transforms them from passive consumers to active participants, and enhances their overall user experience. However, it is crucial to emphasize the importance of explicit consent and transparency when dealing with personal data. Respecting privacy and ensuring that users have control over the information shared is paramount in establishing trust and maintaining ethical practices. 104 105 10. Bibliography “What Is Artificial Intelligence?” (n.d.) IBM. https://www.ibm.com/topics/artificial-intelligence “Definition of Intelligence”, n.d. Merriam-Webster Dictionary https://www.merriam-webster.com/dictionary/intelligence “What is Machine Learning” n.d. IBM https://www.ibm.com/topics/machine-learning Baheti, Pragati (August, 2021) “A Simple Guide to Data Preprocessing in Machine Learning” V7 https://www.v7labs.com/ blog/data-preprocessing-guide#h3 Blue, Juanita, Joan Condell, & Tom Lunney. «Digital Footprints: Your Unique Identity», 2018. https://doi.org/10.14236/ewic/ HCI2018.173. Brown, Sara (April 21, 2021) “Machine learning, explained” MIT. https://mitsloan. mit.edu/ideas-made-to-matter/machine-learning xplained#:~:text=What%20 is%20machine%20learning%3F,to%20 how%20humans%20solve%20problems. Bulchand-Gidumal, J. (2020). “Impact of artificial intelligence in travel, tourism, and hospitality” Handbook of E-Tourism, 1–20, Springer. Claremont Mckenna College (2017) “The Paradox of Choice, Barry Schwartz” [Video]. YouTube, https://www.youtube.com/ watch?v=yhtmhaboE94&t=198s Jiang, Shan. “Exploring Experience and Motivation Using MeansEnd Chain Theory”. Research Gate, November 2019. https://doi. org/10.1079/9781786391896.0053 Morgan, B. (2022, Jun 27). “What Is Bleisure Travel, And How Is It Transforming The Hospitality Industry?” Forbes. https://www.forbes.com/sites/blakemorgan/2022/06/27/what-is-bleisure-travel-and-how-is-it-transforming-the-hospitality-industry/ Papaioannou, Thanos, Aggeliki Tsohou, & Maria Karyda. «Shaping Digital Identities in Social Networks: Data Elements and the Role of Privacy Concerns». Lecture Notes in Computer Science, 2020, 15980. https://doi.org/10.1007/978-3-03042048-2. Schmidt, Eric & Jared Cohen. «The New Digital Age, Reshaping the Future of People, Nations and Business», 2013. https:// doi.org/10.1177/1018529120150208 Uzelac, Aleksandra, & Biserka Cvjetièanin. Digital Culture: Changing Dynamics. Culturelink Joint Publications 12. Zagreb: Institute for International Relations, 2008. Vorobiova, Anna. «The Digital Age: The Era We All Are Living In». Software Engineering. Development Zone, June 2022. https://dzone.com/articles/the-digital-agethe-era-we-all-are-living-in-and-d. What Is Machine Learning (ML)? (June 26, 2020) UC Berkeley. https://ischoolonline. berkeley.edu/blog/what-is-machine-learning/ Kavlakoglu, Eva (May, 2020) “AI vs. Machine Learning vs. Deep Learning vs. Neural Networks: What’s the Difference?” https:// www.ibm.com/cloud/blog/ai-vs-machinelearning-vs-deep-learning-vs-neural-networks Kononova, Olga, Dmitry Prokudin, & Elena Tupikina. «From e-Tourism to Digital Tourism. Terminologically Review». Saint-Petersburg National Research University of Information Technology, Mechanics and Optics, 2020. L. Ceci. “Hours of video uploaded to YouTube every minute 2007-2022”, Statista, 2023. https://www.statista. com/statistics/259477/hours-of-video-uploaded-to-youtube-every-minute/#:~:text=As%20of%20June%20 2022%2C%20more,for%20online%20 video%20has%20grown. Lanckbeen, C. (2023, April 24). “Post-COVID ‘revenge travel’ has gone big. And the revenge is sweet” EuroNews https:// www.euronews.com/2023/04/21/post-covid-revenge-travel-has-gone-big-and-therevenge-is-sweet Levin, Ilya, & Mamlok. «Culture and Society in the Digital Age». Information 2021 12, n. 68 (February 2021). https://doi. org/10.3390/ info12020068 Liu, Yang, & Andrew Simpson. «On the Trade-Off Between Privacy and Utility in Mobile Services: A Qualitative Study». Lecture Notes in Computer Science, 2020, 261-78. https://doi.org/10.1007/9783-030-42048-2_17 Cover Marek Piwnicki, Unsplash, 2021 Fig 1. Social Cut, Unsplash Fig 2. Creative Christians, Unsplash 2020 Fig 3. Card Mapr, Unsplash 2020 Fig 4. Mika Baumeist 2022 Fig 5. Stem T4L Unsplash 2019 Fig 7 .Andrea De Santis, Unsplash 2020 Figure 8. Kavlakoglu, Eva (May, 2020) “AI vs. Machine Learning vs. Deep Learning vs. Neural Networks: What’s the Difference?” Figure 9. Nielsen, Michael (December 2019) “Why are Neural Networks hard to train?” Figure 10. Karnes, KC (2019) “Introduction to (NLP) Figure 11. Baheti. 2021. “A Simple Guide to Data Preprocessing in Machine Learning” Figure 12. Cser, Tamas (2018) Is AI Good for Society? - The Good, The Bad & The Ugly Fig 13 . EV, Unsplash 2018 Fig 14 . Karsten Winegeart, Unsplash , 2021 Fig 15. Cocktail Party Effect, Media Production Lab, EdTech, University of Toronto Fig 16. Fikri Rasyid, Unsplash 2018 Figures 17 & 18. Travel GPT, 2022 Fig 19. iplan.Ai , 2022. Figures 20 & 21. TripNotes, 2022. Figures 22 & 23. The Trip Boutique, 2022. Fig 24. Tom Barret Unsplash 2017 Fig 25. Instagram Saved Content , Travel Files. Fig 26. Instagram liked Content , Travel Related Fig 27. Social Media, Sara Kurfeß, Unsplash 2018. Fig 28. Jason Goodman Unsplash 2019 Fig 29. Generation Y Profile, Inspiring Apps. Fig 30. Generation Z Profile, Inspiring Apps. Fig 31. Interviews insight Fig 32. Brooke Cagle Unsplash Fig 33. User Persona Fig 34. Empathy Map Fig 35. Customer Journey Map Fig 36. Jason Coudriet Unsplash, 2018 Fig 37. Scamper method for ideation Fig 38. Ideation mind map Fig 39. Step by step for AI-powered app design. Fig 34. Business Model Fig 35. App Information Architecture Fig 36. User Flowchart Fig 37. Marco Polo Branding Fig 38. Marco Polo Moodboard Figures 40 through 43. Prototype Fig 44. Faizur Rehman, Unsplash, 2021 Fig. 45 Clay Banks, Unsplash, 2017 106 107 10. Annex Annex 1 – Smart Tourism definitions Annex 2 – Digital Tourism 108 109 Annex 3 – Push & Pull Factors Annex 4 – Travel Career Ladder 110 111 Annex 5 - Different Nationality Profiles Annex 6 - Basic Algorithm # Step 1: Data Collection def collect_user_data(): # Code to collect user data from social media platforms pass # Step 2: Data Preprocessing def preprocess_data(data): # Code to clean and preprocess the collected data pass # Step 3: Profile Creation def create_user_profile(preprocessed_data): # Code to create user profiles based on the preprocessed data pass # Step 4: Feature Extraction def extract_features(user_profile, additional_input): # Code to extract relevant features from the user profile and additional input pass # Step 5: Recommendation Generation def generate_recommendations(features): # Code to generate personalized travel recommendations using machine learning techniques pass # Step 6: Recommendation Refinement def refine_recommendations(recommendations, constraints): # Code to apply filters and constraints to refine the recommendations pass # Step 7: Presentation and Delivery def present_recommendations(recommendations): # Code to present the personalized travel recommendations to the user pass # Step 8: Continuous Learning and Improvement def gather_user_feedback(): # Code to gather user feedback on the recommendations pass # Main Function def main(): # Step 1: Data Collection user_data = collect_user_data() # Step 2: Data Preprocessing preprocessed_data = preprocess_data(user_data) # Step 3: Profile Creation user_profile = create_user_profile(preprocessed_data) # Step 4: Feature Extraction additional_input = get_additional_input() # Code to get additional user input features = extract_features(user_profile, additional_input) # Step 5: Recommendation Generation recommendations = generate_recommendations(features) # Step 6: Recommendation Refinement constraints = get_constraints() # Code to get user constraints refined_recommendations = refine_recommendations(recommendations, constraints) # Step 7: Presentation and Delivery present_recommendations(refined_recommendations) # Step 8: Continuous Learning and Improvement gather_user_feedback() # Execute the main function if __name__ == "__main__": main() 112 113 Annex 7 – Branding Brand Name NARRATIVE Description A travel app that unlocks a personalized world of adventure for every user. With a compass logo symbolizing exploration, the app leverages user preferences and digital trails to curate tailored destination suggestions. Marco Polo empowers users to discover a world made just for them, offering a seamless and intuitive interface for exploring unique travel experiences. From uncovering hidden gems to crafting personalized itineraries, the app embodies the spirit of discovery and provides a trusted companion for unforgettable journeys. Purpose To inspire and empower individuals to embark on personalized and meaningful travel experiences. We believe that travel has the power to broaden horizons, foster cultural understanding, and create lasting memories. Personality Friendly, Resourceful, Reliable and Enthusiastic. If Marco Polo were a person, we would want it to be perceived as a trusted and knowledgeable travel companion with a vibrant and adventurous personality. Marco Polo would be seen as a well-traveled explorer, constantly seeking out new experiences and sharing valuable insights with users. The brand would be perceived as friendly, approachable, and genuinely interested in understanding each users’ unique interests and preferences. Position Because we offer a truly personalized and curated experience like no other. Unlike generic travel apps, we go beyond surface-level recommendations and delve into the nuances of each user's preferences, interests, and digital trail to provide tailored recommendations and suggestions. Our advanced algorithm and intelligent data analysis ensure that every destination and activity suggested by Marco Polo aligns perfectly with the user's unique travel profile. Mission To empower individuals to explore the world in a deeply personalized and meaningful way. We are dedicated to curating tailored travel experiences that align with each user's unique traveler profile. Vision To revolutionize the way people, discover and engage with the world through technology and personalization. We envision a future where travel experiences are no longer generic but tailored to the individual. UNIQUE SELLING POINT Unique Selling Point The integration of social media into personalized suggestions of travel destinations and itineraries. We co-create valuable proposals by partnering with the user who inputs their data. KEY IDEAS Key Terms Honest, trustworthy, authentic, transparent, dependable, innovator. Anti-Terms Generic, limited, cookie-cutter, unreliable, slow, overwhelming. Archetype The Explorer archetype embodies the spirit of adventure, curiosity, and a desire to explore the unknown. It represents a brand that seeks to discover new territories, whether they are physical places or new experiences. MARKET RESEARCH Demographics Digital natives, particularly Gen Z and Millennials (Gen Y), who are accustomed to using technology and value personalized experiences, middle-class, the market is global in nature. Psychographics Adventurous and curious, Individualistic, tech-savvy and connected, experience-oriented, culturally curious, and socially conscious. Competitors Trip it, Trip Advisor, Travel, The Trip Boutique, Trip Notes, Orkoi.