I looked at the following web log analyzers: Webalizer, Analog, and AWStats. I got the first two to work. The installation instructions for the third assume that you’re running your own web server. I’m quite sure one can get it to work on standalone access logs given enough time. But I didn’t try. Analog works fine once the (unusual) format of my access logs is specified to it. It crashed when I turned on reverse DNS to get a more informative summary though. Surprisingly Webalizer was able to read the access logs without being “coached”, although it doesn't recognize the referrer and agent fields correctly. Unfortunately it also doesn’t let you change the format it recognizes so I ended up writing a small Python script to do the format conversion. It’s pretty boring stuff so I won’t post it.
Anyway I’ve generated access summaries for this site for the month of March (minus these last two days) using Webalizer and Analog, respectively. This is the first time I ever have any kind of access information for my site so it’s quite interesting for me. Here’re some observations.
- There are a little over 100 visits per day.
- The Google bots are certainly working very hard.
- Lots of people are feeding off the RSS (but I suspect this is partly due to some people trying to fill up the corners of their web pages).
- I have quite an international audience! (If the visitors with the Niue domain are really from there, please let me know if you need a programmer. I’d love to come!)
- I should really reduce the size and number of graphics on my pages (from the Analog summary).
- There’re big discrepancies in some of the statistics produced by Webalizer and analog. It’s especially strange that there are so few page accesses on the 6th and 11th in the Webalizer report. One must wonder whether its output is in fact correct!
