38

According to official policy, a moderator must have some activity (visit the site at least!) once every six months at a minimum to retain their diamond. I've even seen others lose their diamonds because of this. Meanwhile, other moderators visit just often enough that they keep their diamonds, without doing much else. It is what it is, but it leaves me wondering if 530-something is really the best tally for the number of moderators we have here.

Context: At the time of drafting this (assuming SE is keeping up with the policy), all moderators have visited Stack Exchange since ChatGPT was released to the public, but some have not visited after the new AI policy was announced to moderators, much less during the strike. Despite that, the raw percentage of moderators who signed the strike has been repeated time and time again, even ending up in the media.

While I would therefore appreciate some strike-related data (e.g. how many mods saw the AI announcement), I'm also curious about more general statistics (e.g. how many moderators visit on average during a week/month when there's no strike).

5
  • 9
    how many moderators visit on average during a week/month when there's no strike) Visit tells you almost nothing. The key is measurable actions - flags handled, etc. Which may or may not be available externally, but that is what really matters. Commented Jul 2, 2023 at 14:00
  • 14
    The question is purposefully open ended, since every metric has its downsides. For flags handled, that requires people actually flagging (and your fellow moderators not being selfish and hogging those flags :p) which is tricky on small sites.
    – Laurel
    Commented Jul 2, 2023 at 14:15
  • 4
    I don't think that SE pay very close attention to either visits or moderator actions. Some moderators are basically silent for years
    – Richard
    Commented Jul 2, 2023 at 14:35
  • 1
    @Richard "Some moderators are basically silent for years" Given that the whole network is only 15 years old this is quite stunning in my eyes. Such a moderator should really not count as active and maybe not even as a moderator. I mean how do you know that somebody will come back if he/she wasn't there for such a long time. A sabatical might be ok, AWOL for years... Commented Jul 2, 2023 at 20:02
  • @NoDataDumpNoContribution It does happen, though. A lot of people will become involved in their subject's site during undergrad, get elected towards the end, then go to grad school and have no time to participate, and then come back and become active again afterwards. Commented Jul 4, 2023 at 17:20

2 Answers 2

9

It depends a bit on how you define "active" but if you can live with a moderators "last_access_date" then I have a solution.

The Stack Exchange API (free for as long as it lasts) offers the /users/moderators endpoint. Its result is a user type that has the mentioned last_access_date (yes, that is still the accurate value, not the botched one found in the UI. No guardrails are in place yet).

To get an overview of network-wide activity of moderators you:

  • fetch all sites from /sites
  • filter only site_type = main_site
  • fetch all moderators per site
  • combine all moderators in a single list
  • sort by last_access_date
  • add some UI
  • success

I have done above steps more or less on this Glitch site: se-active-moderator.glitch.me (find its source code here).

The site was up on July, 3rd for the first time (shared in an off-site mod-only space) but it doesn't do timeseries data-collection.

I have made this full page screenshot, shortly before I returned to my site to pick-up moderation again. You can find my name around June, 2nd.

It looks like moderators are always on from this perspective. We don't have better public data.

0
5

Discussing what statistics and details about moderation might be worth showing is interesting. This looks that might derivate in one or more to improve and extend https://stackexchange.com/ features.

Some data about the moderator's activity is available in the per-site community moderator tools. Such data exists, but AFAIK, it's not available publicly. One possible way to proceed is to grab data from https://stackexchange.com, like the moderators' user id, then use the API and SEDE to extract data about their activity first, then aggregate / analyze it. This will not include details like flags handled, but looking at posts and comments might get a good idea of whom is active.


Moderators

I think that there is only one page that consolidate information about moderators.

About > Moderators

https://stackexchange.com/about/moderators

Currently, this page lists the moderators of all sites. By default, it only shows the moderator's user name, but there is a button to toggle details. The details show only the moderator badges summary by type (bronze, silver, gold) and when they were elected.

Someone having good web scraping skills and time might use this page to get details about the mod's activity from their user profile.

Other pages

Pages to consider to improve / extend with statistics about community moderators and moderation activity.

All Sites

https://stackexchange.com/sites

This page lists the sites and shows some statistics, but none are specific about community moderators and their activity.

Top Users

https://stackexchange.com/leagues This page lists the sites and shows the week, month, quarter, year and all-time top user based on the reputation earned in such period. One of the problems of this page is that it shows the top user of the last period that had activity, not the current corresponding period.

Besides reputation, there should be other "leagues" similar to what is shown on the per-community Users page: Voters, Editors. It might be worth discussing the convenience of having a league for community moderators activity, i.e., the number of days active.

About

https://stackexchange.com/about

This page shows statistics, but none are specific about community moderators and their activity.

0

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .