Hey, I'm Mikio and fedistats is my project. I've been building it since January 2023, and I want to talk a bit about what it does and why I built it.
I've always had an interest for social media and ever since Elon Musk took over Twitter last year and I saw people switching over to Mastodon, I thought maybe the time is ripe for an open Twitter alternative. Since then, more people have joined Mastodon, and recently, the total number of signups crossed 10 million.
One big problem is that you cannot just join Mastodon, you need to pick a node. You can migrate your account later on, but you have to make this decision first. There are already plenty of websites out there that crawl node statistics (for example, instances.social), and I had worked with actual live data before and have learned that there are all kinds of interesting trends to be found.
So I started to listen to the federated stream and collect statistics from active nodes. On the Nodes and Languages section of fedistats you can see the live node stats and language stats to help you explore the kinds of nodes that are out there.
On this page, you see the total number of registered accounts, posts, and connected domains as well as the actual number of active users and messages as I'm seeing in the federated stream.
It's not a surprise that mastodon.social is the biggest domain, hosting roughly 1 million accounts, a tenth of all active accounts, as it is the original mastodon site. But second up in terms of daily posts is botsin.space, a mastodon instance for bots and bot allies.
There is also a huge discrepancy between the number of registered accounts and the number of active daily users. I don't know and haven't yet looked into whether the data I get is complete or not, but that's also not that important because I want to see what is happening, and a sample of the data is fine for that as well.
Now if you sort the list by total users (as reported from the sites), the second biggest instance is plasmatrap.com, which is not a mastodon node but runs on Calckey.
Maybe this is a good moment to take a step back to look at the fediverse. Unlike Twitter or Facebook, which are services owned and operated by a single company, the idea of the fediverse was to build a network of interconnected nodes that exchange information via a common protocol, much like Email or the web. There exist several such protocols, but the one that mastodon runs on is called ActivityPub.
ActivityPub by itself only takes care of delivering posts, and managing followers and following, but it is not per se a social network. Mastodon is one project that builds on top of ActivityPub to give an experience similar to Twitter, but there are others. Pixelfed, for example, is more similar to instagram, and so on. Calckey is another Twitter like platform that seems to be quite popular in Asia (at least it looks like from the data).
What is a bit confusing is that Mastodon also provides an API that provides similar functionality (for example, to follow people or post content), but that's an additional layer on top of Activity Pub.
People on the fediverse use #hashtags to mark topics they are talking about. Looking at hashtag can reveal what people are talking about on different nodes to help you understand the community.
For example, mastodon.social's hashtag distribution for the last week looks like this (only the top hashtags shown).
On the other hand, hachyderm.io's hashtags look like this (this is also the node I'm on, after I saw many from my data/ML contact move there).
Completely different hashtags are on infosec.exchange. You can really see the different focus.
For now, if you want to explore nodes on the fedisverse, have a look at the nodes landing page, or at the languages landing page. You can search for node names, look at language distributions, and so on. And if you find something that looks interesting, join the fun!
Leave your comments on LinkedIn, Twitter, Mastodon.
If you want to learn more about my journey and the fediverse, follow me on Twitter or on Mastodon, or follow fedistats.cc account on Twitter or on mastodon.