Weekly Reports: June 6 – June 12, 2016

  • Monday
    • Started running RTTBurst on all of 2015.
      • It’s very slow. About through January 15 after 4 days.
    • Generated a set of misogyny-related tweets
      • Keywords were taken from Hatebase, K. Preston and K. Stanley, “‘What’s the worst thing…?’ gender-directed insults,” Sex Roles, vol. 17, no. 3–4, pp. 209–219, 1987., and S. Hewitt, T. Tiropanis, and C. Bokhove, “The Problem of Identifying Misogynist Language on Twitter (and Other Online Social Spaces),” in Proceedings of the 8th ACM Conference on Web Science, 2016, pp. 333–335.
      • Seems heavy on pornographic material
      • Filtered to remove retweets and non-English tweets
      • Built a CSV sample of tweets for hand-labeling
  • Tuesday
    • Posted terrorism notebooks to github
    • Meeting about SES + Twitter
  • Wednesday
    • Meeting on START relevance project
    • Worked on dissertation
    • Developed a KL-divergence Scala implementation for Spark
      • Available under spark-twitter-nb here
  • Thursday
    • Worked on dissertation
  • Friday
    • Finished a draft of my dissertation intro
  • Saturday
    • Ran KL divergence against Boston, both Paris attacks, Brussels attack, and Nigerian Kano bombing against 1% Twitter at the same time.
      • Top 100 divergent keywords across terror attacks are listed here
    • Built sampling mechanism for Gnip activity data. Will sample a given number of tweets per rule
  • Sunday
    • Hand-labeled 1000 misogynistic tweets to determine distribution of insults, quotes, music lyrics, self-identification, and other types.