For the last year I've been using a plugin for my site that generates a statistics page, through which I can keep an eye on how well the site is doing. A couple of days ago, it stopped working. After a bit of analysis I discovered the data structures it was creating in memory had consumed too much of said memory and my hosting service had auto-assassinated it.
I decided to process the logs myself, by first downloading the logs, then uploading a page with the stats in it. What I found was surprising.
Firstly, I wanted to know the number of unique visitors to the site each month. To be clear, these aren't unique over the entire period, just within each month. What I found exactly gelled with the results I was seeing previously, so that was a good start.
Next, I wanted to know how many of each of the platform-specific versions of Orbital Calculator had been downloaded. While I was testing this part of the script, OC was downloaded again… and then again! I checked the logs to make sure that wasn't an aberration created by the changes in the script, but no, they actually had been downloaded while I was testing. Superb!
In case you're interested, it has been downloaded 204 times (at the time of writing). I can see that all the various platform options have been downloaded in roughly equal amounts, including the Armv7 version, so I know that it's worth keeping all versions on my site.
I next turned my eye to what were the most popular pages. This gives me an insight into what people want and where I can turn my hand to gain the best results. It's a double-edged weapon though, since it can narrow content (and hence the visitor demographic) so I'll always try and maintain the site as a reflection of my diverse interests.
It's not really interesting to see how many hits a page written a year ago has accrued. It's been around longer than a page added last month, so it can distort the analysis. I decided to see what was getting the most views in the current month. I expected this to be the, or the page. It turned out, oddly enough to be , which came top of the list by a large margin.