Observing real world on Twitter

The presentation will consist of results and observations drawn from our survey related to the micro-blogging service Twitter. The motivation behind the survey is to understand better the reasons of users to create tweets, and see if the reasons correspond to real-life situations. We have collected more than 5 million different tweets, by using two different methods. The first one is based on topical keywords, such as "H1N1" and "Olympics". Our other tactic was to gather tweets based on the location of the users, that is, we collected tweets from such cities as Barcelona, Liverpool, Madrid, and Rome.

From the data we have formulated hour-by-hour and weekly distribution of the tweets from different cities. By looking at the distributions we have observed apparent differences in the user behaviour between various cities. The results also indicate that users are tweeting about current events, such as sporting events, awards shows, and topical situations. One observation is that the user are willing to express both their positive and negative thoughts.

We have also done some analysis of the most frequent words used in the tweets, both in the English and Spanish languages. The work is a part of a major study about user-generated content and content distribution.