This demo is the result of a study
that was carried at the Language, Interaction and Computation Laboratory
at the University of Trento in Italy . We looked at the daily patterns of life in Twitter
messages (tweets), and we present the differences in the contents of tweets according to the gender of the users and time of the day.
We analyzed millions of tweets
collected by researchers from the University of Edinburgh between November 2009 and February 2010. For gender differences, we separated the tweets into two subsets as male and female tweets by using the first names of the Twitter users. For hourly differences, we grouped the tweets according to the time of the day they were posted (in each user's local time).
In the gender differences section, you can see which phrases are used more often by males or females. These results are given under the "Compare" tab.
We also looked at the co-occurrence statistics of phrases, seperately for the two genders. That allows us to take a phrase like "cup" and see what else males (or females) talk about they mention "cup" in their tweets in a distinctive fashion (with respect to the other gender). These results are given under the "Detailed Query" tab.
In the hourly differences section, you can see how the frequency of mentionings of a given phrase change over time during the day.
1. Herdağdelen, A. and Baroni, M. (2011), Stereotypical gender actions can be extracted from web text. Journal Of the American Society for Information Science and Technology, 62: n/a. doi: 10.1002/asi.21579