Three Months, Two Spikes, One Conclusion

This post originally appeared on the Software Carpentry website.

Here are the traffic stats for the last three months at software-carpentry.org:

November 2010

December 2010

January 2011

It's hard to make out what's happening because of the big spike in December, and the less prominent (but still significant) spike in November, so here's a summary:

1. Once these spikes are removed, the average number of distinct visitors is slowly increasing each month.
2. Posting a lecture on something popular boosts our readership dramatically, but only briefly, and only for that topic—most of those visitors don't stick around.

So what were the spikes? The first, in November, came when we posted our Python lectures. The second, in December, was Tommy Guy's episode on how to build a simple recommendation engine, which in turn was based on an example from Toby Segaran's excellent book Programming Collective Intelligence. That was more popular than anything else we've ever put on the site, which is both encouraging (if we build it, you will come) and disheartening (I have no idea what else to build that will be that popular).