Work Text:
An anonymous tumblr user asked:
Hi! I really like your stats! I was browsing your character stats published in January 2024, and I was surprised by the fact that Wanda Maximoff didn't appear in the list. I checked, and there are ~38.000 public works under her tag. Did I miss her on the list? In any case, I just wanted to let you know, if it may be useful. Thank you for your great work!
[They were right! You can click here to read my answer (edited slightly for clarity), which explains how this omission and others occurred.]
Hey, thanks for the compliment and for the great question! :) I went back and checked my January big fandom/character/ship stats and also the underlying character data, and you're not wrong! Wanda is missing, and so are others. I'm going to post graphs with data about Wanda plus a bunch more missing characters -- and some missing ships! -- shortly. But here, let me explain how some characters inadvertently got left out.
Until relatively recently, there was no easy way of directly finding the top characters or ships across all of AO3. Now there is a Tag Search feature that lets you sort by usage that allows you to do so (though, as I posted about earlier, it has some limitations). But because I didn't have access to that method until recently, I used to instead do the following:
- Find all fandoms with 10K+ works. (tutorial)
- Combine the top 10 characters (and ships) listed in the Sort & Filter sidebar for each of those big fandoms.
- Find out how many works each of those characters (and ships) have, and make my top characters / top ship lists from that data.
This worked well for many popular characters and ships. However, for really huge fandoms with lots of popular characters, like MCU, it misses a lot of big tags. My hope was that most of the characters that were missed in a giant fandom like MCU would show up in the top 10 list of some other popular fandom. E.g., Pepper Potts isn't in the top 10 characters listed in the Sort & Filter sidebar for the "Marvel Cinematic Universe" tag, but she is in the top 10 for the tag "Iron Man (Movies)." However, Wanda is an example of where this didn't work out. The WandaVision fandom didn't have 10K works, so it didn't make my list of fandoms in step 1 -- and Wanda wasn't in the top 10 for any fandoms with 10K+ works. So I just missed her altogether despite the fact that she appears in more fanworks than other characters in my list.
I will use Tag Search sorted by usage in future years and hopefully avoid this issue. That method would have caught Wanda -- and when I used that method as a starting point earlier this week (and then did follow up work to address the limitations of Tag Search), Wanda came in #108 out of all AO3 character tags! I'll share the resulting data that includes the previously missing characters in a bit. :)
Okay, so let me try to correct this issue! This time, I used AO3 Tag Search to sort character tags in descending order by usage. I then used a scraper to looked at the Works page for each of those tags to find out the actual number of (public) fanworks using the top 250 tags listed there. (See this related ship search discussion where I outline some of the limitations of Tag Search that make this necessary.) The quirks of Tag Search leave some possibility that I'm still missing some characters with a bunch of fanworks, but this method should at least get most of the ones I missed before.
Here's a graph of the character tags with the most works that got left out of my January list of characters with 20+ fanworks (the largest fandom associated with each character tag is shown in italics):
Caption: a graph created from the top 25 items in this longer list of data
Note that a lot of these characters weren't entirely missing from the January analysis -- they were just missing in this particular form (this tag). The character tags with asterisks are ones that did show up in the January list in a different form. For instance, "Thor" wasn't in the list, but "Thor (Marvel)" was. "Thor" contains "Thor (Marvel)" and other Thor-related tags, like "Thor (Stargate)" and "Thor (Phineas and Ferb)." So the above tag is not really a single character, but an amalgam of characters named Thor. It feels arguably okay to have left those big, ambiguous tags out of an analysis of the biggest characters on AO3... but OTOH, there are currently over 10K works in the MCU fandom that use the generic "Thor" tag rather than the "Thor (Marvel)" tag. So by ignoring the more general "Thor" tag, we're throwing those out and therefore presumably undercounting how often people are writing about the MCU Thor character. (There's no perfect answer here.)
(I also said that a tag related to "You" was in the January stats, because the "Reader" tag was in that list -- but these tags are only conceptually related and are not actually wrangled together as synonyms in AO3 currently. As you'll see in the graph below, "Reader" is more common than "You.")
I created new top character and top ship lists that should include all the omitted tags. It turns out that no characters were omitted from the top 25; the only changes in this graph are due to new works (and newly deleted works) since January.
Caption: a graph created from the top 25 items in this longer list of data
What about ships? The same issues that caused me to miss some characters also caused me to miss some ships in my January list of ships with 5K+ works. Here are the biggest ships that were missing from that analysis (once again with the largest fandom tag associated with each ship shown in italics):
Caption: a graph created from the top 25 items in this longer list of data
In contrast to the missing characters, here we see only one ship that appeared in the January analysis in a different form: "Merlin/Arthur Pendragon (Merlin)" appeared in the January stats, which is why the more generic ship tag, "Merlin/Arthur Pendragon," has an asterisk next to it above. This tag is an amalgamation of other Merlin/Arthur pairings, and not strictly limited to the BBC TV show. (Though my January analysis presumably undercounted how often people were writing about the BBC Merlin/Arthur ship, because not everyone in the Merlin (TV) fandom uses the more specific ship tag -- there are over 5K uses of the more generic ship tag in the Merlin fandom.)
IIRC, I may have purposefully left out some of the ship tags that don't refer to specific ships from my January list -- like "Minor or Background Relationship(s)."
There are also some ships on here that weren't listed in January because their fandom didn't yet have 10K fanworks -- including "Alex Claremont-Diaz/Henry Fox-Mountchristen-Windsor" (the fandom tag "Red White & Royal Blue - Casey McQuiston" surpassed 10K works in 2024, though the movie fandom tag hasn't done so yet). And some ships whose fandom has not yet reached 10K public works -- including 'Tyrannus Basilton "Baz" Pitch/Simon Snow' and "Patrick Brewer/David Rose" (both the Simon Snow fandom and Schitt's Creek fandom have over 8K public works currently). Next year, I won't have any such prereqs about fandom size, so this shouldn't block these ships from inclusion in the future.
I once again generated a new list of the very biggest ships overall, incorporating omitted ships. There are very few changes from the January data in the top 25:
Caption: a graph created from the top 25 items in this longer list of data
Merlin/Arthur is higher in the list, for reasons discussed above. We also see the appearance of a few of the non-specific tags like "Minor or Background Relationship(s)." And we see organic changes -- e.g., "Original Female Character(s)/Original Male Character(s)" has grown since January and now makes the list.
I hope this was more helpful/elucidating/interesting than confusing! It was definitely a bit more in the weeds than some of my data. :)