Table 1.
Four example URLs which all refer to the same publication.
Fig 1.
Overview on the different aspects of the analysis.
Table 2.
Basic statistics of the three tweet collections.
Fig 2.
The complementary cumulative distribution functions of the number of tweets per user and the number of URLs per user during 2014 for the URL tweets of the 6,271 computer scientists (CS) and the 5,646 sample users (S).
The “tweets per user” curves show the corresponding distributions for all tweets of the 6,694 computer scientists and sample users, respectively.
Fig 3.
The percentage of users that was active during a specific day of the year (left) and a specific hour of the day (right) for the computer scientists (CS) and sample (S) datasets.
The times were normalized by regarding the time zones of the users from their Twitter profile, if they were available (around 60% of all users have a time zone set in both datasets), else the users were ignored.
Table 3.
The top 15 TLDs for the computer scientists dataset.
Table 4.
The top 20 domains for the computer scientists dataset and for the sample, ordered by the number of users.
Table 5.
The top 20 domains and hosts from the computer scientists dataset, ordered by the odds ratio.
Table 6.
The top 20 publisher domains (by the number of users) for both the computer scientists dataset and the sample dataset.
Table 7.
The top publications from the computer scientists dataset.
Fig 4.
A visualization of the frequent subdomains, host names, and paths for the domain mit.edu.
All subdomains and paths which were contained in URLs that were tweeted by at least 5% of the computer scientists that had tweeted a URL to mit.edu are shown. The numbers give the corresponding percentages of users.
Fig 5.
The cumulative number of tweets over the year 2014 for a selection of publications.
Each publication is identified by its id in Table 7.