Skip to content

Analytics

LiliyaA edited this page May 31, 2016 · 5 revisions

Ideas

  • build a undirected weighted hashtag graph
  • visualize this graph for a given hashtag with depth 3
  • build a weighted directed user graph
  • userA mentioned userB userA: "hello @userB"
  • weight is number of mentions
  • visualize user graph for a given user with depth 3
  • for a given time range: calculate the number of tweets per user
  • cluster places w.r.t. their top X hash tags over a certain time range
  • Do clusters change with time? Is there any temporal regularity?
  • How do clusters depend on the number of the top hash tags?
  • Is it possible to predict top hash tags for a place, a user, a user group?
  • Are there trend setters and trend followers? (among places and people)

in-work

  • for each hashtag calculate the most used neighbours (#79)
  • two hashtags are neighbours if both are mentioned in the same tweet
  • which user spans the largest area with his tweets (#36)
  • calculate the traveled distance between all pairs of subsequent tweets of a user (#78)
  • Which user has the largest accumulated distance?
Clone this wiki locally