Because the title question leads with "How is our tagging situation?" I'm going to start by addressing number 3. If you'd like to skip the detailed breakdown, feel free to jump down to the Comparison Summary heading.
How does our current tagging situation compare to similar sites like Astronomy SE?
To analyze the tagging behavior of Space Exploration in context, I compare it to other Stack Exchange communities of similar age. I include communities +/-2 months of Space Exploration's site age. Those are Network Engineering, Open Data, Freelancing, Blender, Astronomy, and Tor.
I'll start with questions per day compared to number of tags. All questions per day are based on the average over the last two weeks as reported on the network stats. All tag counts are based on a Stack Exchange Data Explorer query.
Site Q/day Tag Count
Space Exploration 5.0 1096
Network Engineering 6.8 447
Open Data 1.5 298
Freelancing 0.2 118
Blender 64 303
Astronomy 5.6 478
Tor 1.3 209
Already you can see a stark difference. Even Blender, which has a vastly higher rate of questions per day (yes, that's 64, not 6.4), has a number of tags within the 100-500 range, just as the other comparison sites, but Space Exploration has a number of tags over a thousand! Now, you might quibble that some of these topics don't have the same number of concepts to create tags for, but if Astronomy created tags for every exoplanet catalogued it would vastly outnumber Space Exploration's tags.
Now let's take a look at how well each site uses their tags. Let's start by sampling one tag from each site. We'll exclude the top ten popular tags and then select the next most popular.
Site Tag % of questions tagged
Space Exploration crewed-spaceflight 3.77
Network Engineering wireless 4.23
Open Data nlp 2.52
Freelancing invoices 4.89
Blender uv 4.68
Astronomy gravity 4.74
Tor onion-routing 4.07
You can see that with the exception of Open Data, every other comparison site's #11 tag is in the 4% range, while Space Exploration falls short. The outlier, Open Data, only has 5,312 questions and 298 tags. Space Exploration, with its question count of 13,681, is more comparable to Network Engineering, which has 13,885 questions and the second highest number of tags of the comparison sites (447 compared to Open Data's 298).
There are at least 222 questions on Space Exploration that talk about crewed missions but don't use the crewed-spaceflight tag, and many (if not most) of those have less than the max number of tags.
To get a bigger picture look, let's compare the number of singleton tags by site:
Site Num of Singleton Tags
Space Exploration 79
Network Engineering 24
Open Data 23
It should be clear by now that Space Exploration does an exceptionally poor job of using tags after creating them.
Space Exploration creates significantly more tags than sites of a similar age, and it is substantially worse at using those tags than those same sites.
What else can we do to improve our "tag hygiene"?
Let's start with why this is a problem. First, I'll address our gross overpopulation of tags. Every time I go to check in on the status of our singleton tags, I find that the majority of the tags didn't even need to exist at all. I remove the obvious offenders, but some of them are more questionable for a mod to unilaterally get rid of. If everyone followed the guidance at Should I create a new tag?, then we wouldn't have an outsize number of tags to begin with.
Next is tag usage. Under-utilization of tags is a big problem for a few reasons:
- If tags are not being used, why keep them around? We have a load of tags to maintain for no discernible purpose.
- Tags help categorize questions. Categories make searching for existing knowledge a lot easier. We would get less redundancy in new questions if people could more easily find the information that's already available. Tags help with that. If hundreds of questions exist about a topic but aren't tagged with the corresponding tag (as with crewed-spaceflight), then those questions may get lost in the sea of search results.
- User maintenance of tags partially depends on how much a tag is used. Creating tag synonyms is a big part of tag maintenance, and it can only be done by mods and users with a certain score on the tag. If tags are used on less questions than they could be, then less users are getting the opportunity to be able to help maintain that tag.
So that leads me to what we can do about it. It is impossible to control the tags that each new question asker applies initially, though anyone reading this should keep the meta guidance on tagging linked above in mind when asking new questions. What we can control is the review process. I know a lot of us are involved in editing new questions or reviewing first posts on the review queue. When we do that, we need to keep an eye out for what's going on in the tags.
What are good strategies to check the tagging of posts from new users or posts we are already editing for other reasons?
It is really easy when reading a question to take a moment to mouse over the tags and see if any of them are new. If you see anything that says "1 question" in the pop-up box, it is a new tag. Really, it's probably best to consider anything used on only two questions as a new tag, as well.
If the tag is new, walk through the steps of our meta guidance on tagging. Remove bad tags, and if necessary replace them with appropriate tags. Also, make sure the tag has a good excerpt, because people are more likely to reuse it if there is a brief explanation of what it's for.
When removing bad tags (new or not), if the tag is used on two or three questions, make sure to remove/replace the tag on all of the questions. If it's used on more than that, it's probably better to bring it up on meta rather than edit a bunch of questions unilaterally.
In addition to monitoring new tags, you should also look over the other tags to make sure they are appropriate for the question. When doing this, consider what the essential subject matter of the question is and what tags relate best to that.
If the question only has one or two tags, consider adding a few more. Look up words connected to the subject matter on the Tags page. Use Google to find synonyms and related concepts if it helps.
I have some queries at Stack Exchange Data Explorer (SEDE) that you can use to help monitor tags, if you like. Keep in mind that SEDE is updated weekly on Sunday, so you can't see the latest activity there.
Over a thousand tags is a lot, but it is still within our capacity to shape up our tagging behavior. The site is still growing, and we want to catch this now before it really balloons into something unmanageable. If you see a tag you have questions about, ask about it on meta. We need to be having conversations about these tags, so that we can determine as a community how to get the best use out of our tags.