The year in log files

For a little more than four years now I have kept half an eye on visitors to my various sites using Jason McIntosh’s Bise. I like the hand-rolled ritual each Sunday morning of renaming the log files, running my duct-taped scripts, the brief shot of elation (or despair), and then putting the log files away, never to be looked at again. Except, I do look at them again occasionally, and this is one of those occasions.

A couple of years ago I came across Tim Bray’s post about Topfew and had a play with that myself. A primary insight from that little exercise is that my most popular posts by far were targets for spammers who won’t take no for an answer, which I could do nothing about except practice more frequent personal hygiene. Anyway, I just ran the “old” topfew on the 2024 logs from four sites of mine, and here is what I found.

This site, the Mothership; ~3.1 million records, 0:27 to analyse

Most popular destinations are the various flavours of feed and the list of most recent blog posts; that is very gratifying. Among individual posts, number one was a strange little recent post about getting help from ChatGPT to solve a very trivial automation. Must have struck a chord. Good to see, too, that the spammers are still knocking like crazy on WordPress’s well-known doors, even though this site is nothing to do with WordPress. Might be helping in its own teeny way to keep them out of mischief elsewhere.

Fornacalia; ~ 1.7 million records, 0:16 to analyse

This is a WordPress site and the top destination was indeed one of those immortal posts that, for some reason, spammers cannot keep away from. I say “for some reason” because no spam post has ever been published to that or any other post. So why do they keep coming? Other than that, a lot of the top destinations are to the /embed form of the URL. Are these coming from places where the post has beeen shared? I’m not sure. Top actual post is my recreation of the ancient Roman panis quadratus, also gratifying.

My Known microblog; ~ 15 million records, 4:09 to analyse

Most data, least information. Top hit is the RSS feed, probably because both this site and micro.blog pull it in. Beyond that, it is hard for me to make sense of what is going on, just as it is hard for me to make sense of WithKnown behind the scenes.

Eat This Podcast; 9.4 million records; 1:28 to analyse

Confirmed again that the same old posts are attracting the same old spammers, and that /embed is popular too. I’m rather pleased that the episode on folic acid is the top post, because it was one that I put a lot of effort into and that I think addressed an important topic.

Learnings

Not many, except that I do enjoy being able to take a high-level view of what is going on. Also that I need to do even more with the four years of accumulated data. Not that it will affect what I actually do, because chasing an audience is the surest way to lose yourself. And in the end, gratitude to Jason McIntosh and Tim Bray for providing toys to amuse myself.