Gem Download Study: The first 12 weeks.

As most of the readers of this blog know, I release a lot of Ruby language gems. They are announced on this web site regularly. Now it is possible for me to check how often these gems are downloaded. You can see that data here.

Now fairly early on, the view was expressed that most (if not all) of downloads I was seeing were the result of “bots” that grab everything they can find and gather the results for their industrial overlords, most likely in China. For some time I’ve wondered if that was true. To be honest, I hope it isn’t, but the nagging doubt never left me.

So I came up with a plan. I would track the downloads of my gems on a weekly basis for 12 weeks and see if I could see some trends that might indicate the nature of the download traffic. Well, it’s been 12 weeks and here are my initial results:

Gem_Downloads_12Now I know it’s a bit of an eye chart, but just click on it to get a better view of the data. Maybe even open it in another tab for easier reference.

The first thing that becomes visible is that the slopes of the lines, and thus the rate of downloads, is not the same for all of the gems. Some of the gems are clearly more popular than others.

Even more telling, is that for one gem, “vls”, a change was made recently, that described the use of this gem with the popular “rails” web framework/library/system. Once this appeal to this very active group was made, the rate clearly began an sharp upward trajectory.

Now this is only 12 weeks worth of data, so one should not read too much into these results, but they seem to indicate that at least some of these downloads can be traced to real people. Time will tell. I plan on my next study release at the 24 week point.

Yours Truly

Peter Camilleri (aka Squidly Jones)