Big Data Text Analysis

Social Media Research with Word Association Thematic Analysis
Date: Pick your own day and try to follow the suggested timings below
Instructor: Mike Thelwall m.thelwall@wlv.ac.uk

Before the day of the course:

(1) You will need a Windows computer to download and run a program from. Download Mozdeh from the Download link above and save it to your computer or USB stick in a location, such as My Documents, where you can create folders. Mozdeh is free academic software with no malware. Double-click on Mozdeh after downloading to make sure that it starts without getting a computer error (click the Test button in the Project Wizard screen and then Quit afterwards).

Possible problems and solutions: (a) if you get a file access error and are running Mozdeh from a network drive, the cloud or storage not on your computer, please try running it from a USB stick attached to your computer because network delays can cause it problems. (b) If you get an error message about proxy permissions when running Mozdeh from a work computer, then please either get your network administrator to allow Mozdeh to access the internet or run it from a non-work computer, such as from home, if you can.

(2) THE TWITTER PART NO LONGER WORKS - BUT attendees were previously asked to sign up for a free Twitter account, if not already a tweeter (follow @Mozdeh4 for updates). Also, sign up here for a Twitter API developer account and take a copy of the huge "bearer token" that will be assigned to you. Mozdeh will ask for the bearer token.

(3) Please learn how to email screenshots in case of technical problems during the course.

(4) You will also need a program that can open Excel spreadsheets, such as Excel or Open Office.

(5) If interested in YouTube, get a YouTube API key with the Workshop Appendix instructions (about 1 hour), saving the key somewhere safe.
Optional:  Two talks about research projects made with the help of the methods in this course. ADHD tweets, Bullying on YouTube [Bonus video about Word Association Thematic Analysis].

Click to download Zipfile of slides and workshop tasks. Note that a PIN is no longer needed, but a bearer token is (see above).

10.30	Talk 1: Introduction to Word Association Thematic Analysis: discussion of some WATA papers, overview of WATA, and introduction to context and theme detection [YouTube talk]
11.20	Coffee break
11.40	Workshop 1: Practice word association thematic analysis coding. [Starting with a live demo. Practice talk introduction; Demo; THE TWITTER PART NO LONGER WORKS]
12.30	Lunch
1.30	Talk 2: Word association mining with Mozdeh. [YouTube talk][Mozdeh demo of searching for Tweets THE TWITTER PART NO LONGER WORKS].
2.20	Workshop 2: Downloading tweets and saving WATA information with Mozdeh [WATA demo in Mozdeh] THE TWITTER PART NO LONGER WORKS
3.10	Coffee break
3.30	Talk 3: Mozdeh exporting WATA information to a spreadsheet. [YouTube talk]
4.20	Workshop 3: Exporting WATA information to a spreadsheet with Mozdeh. [demo]
4.50-5	Talk 4: Summary and next steps. [YouTube talk] Finish.

Some spreadsheets for Workshop 1:

• a) Words associating with journal articles about men vs. words associating with journal articles about women (small example, only three words). [blank for practice:WataExerciseJournalArticlesManVsRestSmallBlank] and [completed:WataExerciseJournalArticlesManVsRestSmall]
• b) Words associating with bullying comments vs. words associating with other comments from a set of YouTube influencer video comments.[blank for practice: WataExerciseBullyingVsRestBlank] and [completed: WataExerciseBullyingVsRest]
• c) Words associating with early Covid-19 Tweets vs words asssociating with later Covid-19 tweets. [blank for practice:WataExerciseEarlyLateCovidBlank] and [completed:WataExerciseEarlyLateCovid]
• Words associating with high vs. low retweeted tweets from a set of ADHD tweets.[blank for practice:WataExerciseHighLowRetweetsADHDBlank] and [completed:WataExerciseHighLowRetweetsADHD]
• Words associating with the UK vs. the USA from a set of #BlackLivesMatter tweets.[blank for practice:WataExerciseUkUsaBlmBlank] and [completed:WataExerciseUkUsaBlm]
• Words associating with male tweets vs. words associating with female tweets from a set of March 2020 Covid-19 tweets.[blank for practice:WataExerciseGenderCovid10-23March2020Blank] and [completed:WataExerciseGenderCovid10-23March2020]

See also an overview of WATA with Mozdeh.

Made by the University of Wolverhampton during the CREEN and CyberEmotions EU projects and updated at the University of Sheffield.