search

bellecarrell / twitter_brand

In developing a brand on Twitter (and social media in general), how does what you say and how you say it correspond to positive results (more followers, for example)?
0 stars 1 forks source link

07/30 meeting #8

Closed bellecarrell closed 6 years ago

bellecarrell commented 6 years ago
  1. ideas for paper
    1. use another paper's method for extracting data (paper, code) maybe more relevant
    2. use static (not changing over time) factors for success (for ease) at user and tweet level, including log success from Cohn paper http://www.aclweb.org/anthology/E14-1043 and # reweets or something more complicated
    3. current research questions: does staying on topic about brand correlate with success? does current status (# followers) affect advice, or does it change newbies vs. larger influencers? thinking of bringing in dirk hovy to see if marketing research has insights usually applied to corporations that align with or refine these questions
  2. log success for different influence levels for "designer" role in email
  3. troubleshooting on #7
  4. your future involvement
bellecarrell commented 6 years ago
    1. yeah, use
    2. time series analysis more interesting, Cohn success metric modified (delta in/out)
    3. topic model distribution (KL single vs. history), conditional prob that word is generated; Shawndra Hill
  2. instead of bumping up to next level, analyze differences in quartile ranges (people accruing followers the most in range)
  3. pickle load option on python3?
  4. free for part of august