Is Facebook the new kettle?

The new series of I’m A Celebrity, Get Me Out Of Here (or IACGMOOH to those that like acronymns) kicked off on ITV on Sunday night. As is the case with many things in this day and age, there are a number of ways to fly the flag for the programme above and beyond just watching it.

One of those ways is via being a fan of the IACGMOOH on Facebook. If you want to be a fan of the programme on FB then it is just a matter of clicking on the like button at the celebrity.itv.com site and you can let all your friends know of your allegiance to IACGMOOH.

Likewise, you can also “like” any of the contestants in the show by clicking a “like” button on their individual page.

And now the technical bit…
I was curious as to how the number of fans of the show changed through the course of the programme and so set out to first collect and then visualize the data.

The Facebook Graph API (documentation here) offers some very easy to use http-based queries to get hold of data about such things (as well as status updates where users have fairly lax privacy settings, but more about that another time). For example, try clicking on this link: http://graph.facebook.com/search?q=football to get the latest 25 publicly available status updates mentioning football.

In the case of being a fan, the query is of the form http://graph.facebook.com/?ids=nameoffanpage which returns some JSON data. In the case of celebrity the ids is “Imacelebrity”. For any other page, go the page and look at the URL in the address bar to get the page name.

To collect and turn this data into something more tabular, I used R (and particularly the RCurl and XML libraries) to parse the data and then stored it in a MySQL database. The script ran once a minute for the duration of the show to build up a time series dataset.

So what happened?





Looking at the data for the first show then, the average number of incremental fans per minute (fpm) during the 90 minutes of the show + 10 minutes after was 185 fans, with a peak of 323 fpm at 22:32. There seems to be peaks during the programme breaks for the programme, with the biggest spike occurring at the end of the programme – rather than putting the kettle on, ad breaks and the end of the programme seem to be the time to go online.