State of the Cafe, Year Eight

I’m going to do a relatively quick info-dump, and then maybe follow up with some additional analysis and commentary later.

I may come back and just do 2024-2025, but I thought it’d be interesting to look at our entire history first:

Consolidated Page Views

Daily Engaged Users

Daily Active Users / Monthly Active Users

Likes

Posts

Topics

Trending Search Terms

All date ranges are 2017-04-20 to 2025-04-20, and all data is included with the exception of Trending Search Terms which I’ve truncated to the top 20.

Feel free to ask any questions, and suggestions are welcome for better ways to present the data.

37 Likes

The last image, “trending search terms”, has had me laughing out loud for the last five minutes.

The fact that “penguin” ranks three places above “elon musk” is, what’s the right word?
Serendipitous maybe?

Anyway, thanks for the chuckle, I needed that.
:grinning:

39 Likes

What the heck was going on in January/February 2021?

28 Likes

This is probably the data that lends itself least to this time range. It would likely produce more interesting results over shorter time ranges like yearly or even monthly. I don’t have any immediate ideas on how to display that concisely. I’m not opposed to running the report 96 times, but I’m not sure anyone wants to see all 96 of them…

Unknown. My best guess is that there was an overall spike in Crawler activity, and the traffic was misidentified as anonymous. Here are a few snapshots of the data that highlight this:


For two days in November 2020, there was a spike of 10-20k in anonymous pageviews, while there’s a relatively small uptick in crawlers.

Here’s a smaller spike approximately a month later.

Here are a couple of large spikes around 10 days apart, including the large one you pointed out. I have no idea why the magnitude of these is so large. I certainly don’t remember anything happening around those times. There’s not a corresponding spike in posts or topics either, so that doesn’t explain it. Ignoring magnitude for a moment, I think the general pattern is that crawlers are frequently doing small or incremental updates, with periodic large or full updates.
Here are a few more snapshots for context, although none of them are quite so dramatic:


This last one shows where we’re starting to see logged in pageviews eclipsing both anonymous and crawler, sometimes even more than both combined.

24 Likes

Here’s one I missed in my initial data dump:

It seems like predominantly search engines, plus a couple of mail services at the bottom, and then a very specific one right in the middle.

This could be another one that could look interesting over time, rather than in aggregate.

28 Likes

Elsewhere was used by a bunch of people in some sort of game at one point. Does that have anything to do with the odd spikes?

19 Likes

20k+ pageviews in a single day is still high, even for that. If they were bots, then I would expect those numbers to show under logged in pageviews instead. The logged in counts during those spikes is not out of line with the other days around them, so there’s not a similar increase in traffic, which is part of why those days stand out so much on the graph. That said, I’m not familiar enough with how the games were being run to be able to definitively rule it out either. I would defer to @tinoesroho’s expertise there.

22 Likes

Golly I wonder what happened here?

31 Likes

image

27 Likes

It is a steep incline, isn’t it?

Here’s what the actual numbers looked like:

28 Likes

And here it is:


Yellow (gold?) is crawlers, red is anonymous, and blue is logged in.
No surprises here. If you’re wondering why you occasionally see errors/performance issues with the site, this is why.

Roughly and 3-4x increase of engaged users.

From ~50% to ~75%.

It’s not obvious from the graph since the y-axis is basically broken, but there were regularly dozens of likes. It’s just that now there are thousands.

It’s not an accident that this graph looks almost identical to the one above it.

The userbase here has never been much into creating lots of new topics, instead preferring to keep to relatively few. It’s not obvious from the graph because of the how the data is exported, but there are actually many gaps in the data, since it only contains dates when new topics were created.

30 Likes

Needs a logarithmic y-axis?

17 Likes

I’m very curious as to what sort of game they were playing.
Any idea how it was played?

21 Likes

Calvin Ball; usual rules apply.

25 Likes

They were role playing games, in the play by post style (more or less.) The first game we played was set in the world of Highlander, and the second was called Space Abbey, which was a sort of futuristic space opera/Downton Abbey hybrid… and they were wonderful fun! I was always amazed at what kind of amazing stories my fellow players came up with.

They’re still up on our site, and I believe most if not all threads involved are tagged with badass (the system used) and/or the Games category. @Messana used to run them, but they haven’t been around for a while. :cry:

31 Likes

I said it before but doesn’t hurt to say it again: thanks for your service @LockeCJ

35 Likes

I always used to wonder what the “badass” tag was for.
They sound like fun. Would I find it if I scrolled through the games thread?

and

:rofl:

24 Likes

Here you go:

I get that it makes mathematical sense, but I find log scale graphs to be difficult to parse mentally.

26 Likes

I can do better than that. Here are the starter posts for both games:

I’m not sure how much sense they make to non-players; I once tried to read an old game at TOS and quickly got confused, but I think that was before I started playing the games here. And for each game there were a number of DM threads, some with mechanics and some with player and character interactions that can’t be shared.

25 Likes

Hi
One of the challenges I’m having as a newb here is it’s difficult to wade through the many posts.

Too many choices - like surfing cable listings.

There’s utility in having a handful of ‘contributors’ - like TOS - writing short articles on a main page. I find it much more approachable.

No idea if it’s technically doable or even broadly desirable. Thx
Ps - plug for more reaction emojis!! :pray:

21 Likes