Introduction to corpus building and text analysis with Voyant Tools
Abstract
The slides published here were created for teaching basic text analysis / distant reading with Voyant Tools to MA students in the humanities and social sciences. The focus is on analysing social media data. When the course "Machines of Knowledge" at Maastricht University was first taught, we used Twitter for data collection. After Elon Musk's take-over of the platform and changes to the API, we began using Apple podcast reviews and YouTube comments instead. Also, the technologies we use for data collection have evolved over time. Initially, we used browser-based scraping tools such as Netlytic, given that most of our students have no technical background. In the meantime, we are using Python code that students can easily adjust to collect their own data. The developments of the course are reflected in the slides from the different academic years.
Full text
Text Analysis 2: Voyant Tools Machines of Knowledge - Maastricht University - November 2022
Ingesting data into Voyant: technical issues?
Overview: what are the "tools" in Voyant?
Guide to all tools: voyant-tools.org/docs/
Types of tools: graphs / visualisations versus grids / tables quick overview of distributions / trends spotting co-occurrences and relations understanding networks better access to numerical information more detailed comparison of data easier (de-)selection of words
How to add new tools to your panel:
Avoid overly complex or animated tools! Not all tools, especially animated tools like the "knots", are easy to interpret or suitable for inclusion in written works such as your essays!
Tools have different names - but may be the same! Tools such as "trends" and "stream graph" look different but serve the same or similar purposes.
Don't trust what you see! Some tools such as the "Dream Scape" maps are not fully developed and may contain errors or mislead you. In many cases, they will simply be "empty".
Voyant in practice: analysing reactions to Elon Musk's 2022 Twitter takeover
High-level analysis with word cloud and frequencies table Comparative analysis with the "trends" tool Analysing co-occurences in the terms berry and/or other "links" tools Reading keywords in context Drawing general conclusions Tasks to perform in Voyant Tools:
1st step: open the task sheet on Github & ingest the data into Voyant Tools (using the GitHub URL) https://github.com/MonikaBarget/DistantReading/blob/main/Elon Musk_task-sheet.md
2nd step: updating stop words (OPTIONAL) You may want to add very frequent words temporarily to make working with the word cloud easier.
Step 3: start your analysis with the word cloud...
People whom you may not have been able to identify... "L'institut Apollon fondé et dirigé par Jean Messiha a une ambition affirmée : préparer le basculement en 2022 de la France vers le « camp de la France »." Vijaya Gadde is an American attorney, who served as general counsel and the head of legal, policy, and trust at Twitter. Ned Segal is an American business executive. He was the chief financial officer of Twitter from 2017 to 2022. He was fired following Elon Musk’s purchase of the company. Parag Agrawal is an Indian-American software engineer who was the CEO of Twitter, Inc. from November 2021 to October 2022. He was fired on October 27, 2022.
Surprising elements: frequent mentions of NFTs and cryptocyrrencies (Bitcoin & Dogecoin) "Tesla CEO and Twitter chief Elon Musk has made bullish statements about bitcoin and dogecoin despite crypto market sell-offs. He said bitcoin “will make it” and “DOGE to the moon.” Amid crypto winter and the chaos surrounding bankrupt crypto exchange FTX, Musk believes there is a future for bitcoin, ethereum, and dogecoin." (news.bitcoin.com) "Former first lady Melania Trump is back in the NFT game, auctioning a new set of assets that commemorates moments from husband Donald Trump's presidency." (CBS News)
Step 4: comparative analysis in "trends": people
Step 4: comparative analysis in "trends": countries
Step 5: Analysing cooccurrences with terms berry, word tree, links or correlation tools
Any questions?
tomorrow's lecture by Susan will introduce you to reading data from a postcolonial perspective prepare a text analysis in small groups, using the Amanda Gorman, the Qatar or the Oslo data sets To prepare...
[email protected] Thanks for your participation!