These are the old pages from the weblog as they were published at Cornell. Visit for up-to-date entries.

March 19, 2004

Feed Analysis: Weekday vs. Weekend

Do we read our feeds differently during the week then in the weekend? In the two months that we are investigating, the weekends received about 85% of the pull requests compared to the days during the week. This first figure shows the distribution of the number of pull requests over the days of the week.

Given that we are interested in pull frequency and user behavior we look at the number of requests per day that a user makes and compare the histogram of Monday through  Friday with that of the weekend days. Below are two graphs; the first compares the histograms of the absolute numbers of request (normalized to 100). Here we see that for most categories the weekends indeed run at about 80% of the weekdays, except for the 24 requests per day category, which is about 23% more than during the week. I believe this increase is an artifact of the way that I assign the notion of a unique user, during the weekend these users (laptop?) do not change IP address, which is something they will do during the week when carried back and forth from home to work.

The second graph shows the behavior in a relative sense, and we see that except for the 24 requests per day category there is no significant difference between week and weekend behavior.

Posted by Werner Vogels at March 19, 2004 11:48 AM