Twitter is a social networking service.
Interacting with the API revolves around posting messages and view specific users' messages.
Querying the Twitter API - Both single requests and mass queued up "GET ALL THE TWEETS" - how it was done.
Data Mining - attempts
With this basic grasp of the twitter api, I created the Auto feature which makes multiple GET /statuses/user_timeline to a specific user to download all the user's tweets. Once downloaded these tweets can be analysed in different ways. Here are few of them.
- tweet features
- What hashtags does/did this person use and how often?
- Which users does this person reply to and how often?
- What tweets did this person retweet?
- What twitter clients does this person use and how often?
- Time of day when person tends to post.
- deeper thinking
- frequency of words used in all tweets - which could be turned into a word cloud
- frequency of words used in tweets replying to specific users
- hedonometer - the "happiness" level of a tweet, which can then be analysed over time to track mood changes. http://hedonometer.org/
- identify depressed (clinical / temporary) persons
- could even be used to identify suicidal persons ? http://news.cnet.com/8301-1023_3-57565547-93/facebook-boosts-efforts-on-suicide-prevention/
- readability test - Flesch–Kincaid
- might not be suitable because tweets are limited to 140 chars