Mapping people and tags on Mastodon

In Mastodon relationship graphs I showed how to use Steampipe to map Mastodon network neighborhoods. When I use the word map here, I’m channeling Denis Wood’s The Power of Maps:

Every map shows this … but not that, and every map shows what it shows this way … but not the other way. Not only is this inescapable but it is precisely because of this interested selectivity—this choice of word or sign or aspect of the world to make a point—that the map is enabled to work.

The aspect chosen by those neighborhood maps is the boost—the Mastodon version of a retweet. One of the maps focuses on a selected instance that appears in the home timeline. It shows people who belong to that instance and who boost toots from people on the same or different instances.

The other map zooms out to show boost relationships among all the instances that appear in the home timeline. This view wouldn’t be legible if it included people, so it omits them in order to focus on server-to-server relationships.

These maps represent (or as Denis Wood emphasizes, “re-present”) a set of toots. They omit original toots to which nobody replies, and they also omit replies, in order to focus on boost relationships. What about replies? That would be a different map, one that might also be interesting to draw.

Meanwhile, though, I’ve come up with another map to display the tags that appear in the results of a Mastodon tag search, along with the accounts that use those tags. It proved its worth today when I was looking for perspectives on Section 230 of the Communications Decency Act. As you’ve likely noticed, the US Supreme Court is reconsidering Section 230. My understanding of the topic wasn’t current, I wanted to refresh it, and I especially wanted to test whether Mastodon could provide a useful alternative to a conventional web search.

One affordance that Mastodon provides: search for toots that use the #Section230 tag. Here are two ways to map the results of that search.

IDG

On the left is a conventional Mastodon view: a list of toots that match the query. In this case the article I ultimately wanted to read appears way down in that list. The toot that announced it was from The Markup, “a nonprofit newsroom that investigates how powerful institutions are using technology to change our society.” The article, Section 230 Is a Load-Bearing Wall—Is It Coming Down?, transcribes part of a conversation with two legal scholars whom I know to be reliable guides to Net-related issues.

On the right is my Steampipe-based Mastodon tag explorer. Working with the same data, it surfaced The Markup’s article in a way that brought it immediately to my attention. The first thing that caught my eye was the conjunction of two tags: #section230 and #scotus. Since the Supreme Court’s interest in Section 230 is what’s driving the current news cycle, I wanted to hear from legal scholars qualified to discuss the court’s interest in Section 230. So the tag conjunction was a significant landmark.

The map displayed two nodes that connect to both #section230 and #scotus. How did I choose between them? My prior familiarity with The Markup led me to click on that node and visit the Markup’s Mastodon instance where I read the article.

Had I been following The Markup then, as I am now, I would likely have seen the article on the news list to which I’ve assigned The Markup’s account. But that wouldn’t have changed the experience of searching for the #section230 tag. The relationship graph works by reformulating the results of that search. It omits the text of toots that contain the tag, and the images in those toots, in order to highlight two aspects of the result list: people (or accounts) and tags. It contextualizes those tags by charting their relative frequency in the result list. And it attaches, to each tag node, a link to a new graph focused on that tag.

This “interested selectivity” enables the map to do its work: find accounts that use given tags. Like a tag node, an account node provides a link—in this case, to the account’s Mastodon home page. It also reports the account’s description using a property that appears when hovering the node. So if I were unfamiliar with The Markup I could reveal its description without leaving the graph. Here’s the query that adds that property to the node.

select
  note
from 
  mastodon_search_account
where 
  query = 'https://mastodon.themarkup.org/@themarkup'

+---------------------+
| note                |
+---------------------+
|  Watching Big Tech. |
+---------------------+

That query is embedded in another query that joins across two Steampipe plugins: one that wraps the Mastodon API and another that queries RSS feeds. That’s because, as noted in Mastodon, Steampipe, and RSS, the RSS feeds that Mastodon provides for tags enrich the results available from the core API.

Enabling SQL to query diverse APIs in a common way is one of Steampipe’s core superpowers. Enabling such queries to form the nodes and edges of relationship graphs is another. Used together, these two superpowers enable maps that select what is helpful, omit what isn’t, and thus “re-present” information for a given purpose.

This series:

1. Autonomy, packet size, friction, fanout, and velocity
2. Mastodon, Steampipe, and RSS
3. Browsing the fediverse
4. A Bloomberg terminal for Mastodon
5. Create your own Mastodon UX
6. Lists and people on Mastodon
7. How many people in my Mastodon feed also tweeted today?
8. Instance-qualified Mastodon URLs
9. Mastodon relationship graphs
10. Working with Mastodon lists
11. Images considered harmful (sometimes)
12. Mapping the wider fediverse
13. Protocols, APIs, and conventions
14. News in the fediverse
15. Mapping people and tags in Mastodon
16. Visualizing Mastodon server moderation
17. Mastodon timelines for teams
18. The Mastodon plugin is now available on the Steampipe Hub