Project Goal

The goal of this project is to explore the use of hashtags as a method of social classification on Twitter. We make the assumption that users mark a word with a hashtag as a way of categorizing a tweet. The question is whether there is any consistency in the words people use for hashtags. We attempt to answer this question by getting a bunch of tweets that contain a given word and seeing how people have categorized (hashed) them.

Project Method

We used the Twitter API over a 5 day period to build up a database of around 150,000 tweets. We then created our own API against a MySQL database. When given a word our API does a full text query against all of the tweets. It returns a JSON object containing the count for all of the unique hashtags in tweets containing the given word. The example below does a query on the word obama.
http://abecoffman.com/twitterviz/twitterviz.php?word=obama

Project Results

The three visualizations display the hashtag counts in various ways.