Who borked the Area 51 stats?

Question

All beta sites (e.g. Anime & Manga) are appearing with 0 Q/D and 0 avid users. Site index is unaffected.

In the Q&A spirit, my question is: Who borked the stats?

Answer 1

It was meeeeeeeeeeeee!

A reference change to the underlying libraries didn't expose itself as an issue until the library that relied on it was called. As luck would have it, erroring requests in the multi-page fetcher were also swallowed (not intentionally). Silent fail FTW!

A fix has been deployed and stats should be updating properly now.

Fun fact: I don't have an Area 51 database locally. This fix brought to you by blind luck guesses and pushes to production for testing.

Answer 2

In the meantime, the figures on the site's home page look correct. The Stack Exchange site list also has a correct-looking figure for questions per day.

I wouldn't be sad to see those Area 51 figures go away altogether. If they're going to stay, please at least correct the goals. As it is now, each beta invariably has people who compulsively compare the current figures against the goals and obsess about them, instead of looking at the big picture.

• The “questions per day” figure isn't very useful and many sites graduate without having reached the 15/day goal. How many posts are necessary to sustain a core community is highly variable; sites have survived for years with 1 q/day.
• Percentage answered is kind of useful, but mostly in its evolution on a given site. The figures are very dependent on the topic; for example language sites hover typically around 99% (for pretty much every question, there's a native who has something to say about it) while at the other end of the spectrum sites that involve software products have questions that are only answerable by the very few people who use exactly that same software function.
• Avid users isn't useful in its present form. Active users at certain reputation milestones would be useful as a graduation requirement — but even that shouldn't be displayed publicly as it pushes towards focusing on the metrics instead of focusing on quality (good answers, easy-to-find questions, welcoming ambience).
• Answer ratio isn't nearly as useful as quality checks on the answers. Community evaluations are much better for this.
• Visits per day is a measure of how the rest of the world finds the site useful… it's kind of nice to see that grow, but utterly useless as an absolute metric, and not that important: if you build it well, they will come.