One of the data challenges when it comes to looking at fanfiction--and specifically representation within fandom--that I haven't figured out a good way to account for yet is that it is very hard to capture the popularity distribution of works within fandoms.
What I mean by that is that, while you can see how many fics are tagged in each category (M/M, F/M, F/F, Gen, etc.), for example, you can't easily capture how popular fics in each category are relative to each other.
NCIS shows a clear example of this problem and how it complicates determining what is "popular" in fandom: as of June 10, 2026, there are 8424 Gen fics, 8093 F/M fics, 7483 M/M fics, 1029 F/F fics, 685 fics, and 246 Other fics tagged to NCIS on AO3. From those numbers, you would think that Gen and F/M fics are more popular than M/M fics, because more people are writing them.
However, the most popular fics are virtually all M/M fics. 18 of the top 20 most kudosed fics are M/M, while only 4 of the top 20 are F/M (fics can be multiple categories, and 3/4 top F/M fics are also M/M fics), and 1 of the top 20 is Gen. This holds fairly consistent for at least the top 100 most kudosed fics in the fandom. So based on what fics are popular, M/M is by far more popular than the other categories in the fandom. People respond more positively to them; they do better in the fandom.
Unfortunately, kudos distribution is basically impossible to collect at scale. Unlike category, which is collated by AO3, kudos data has to be gathered entirely manually, which means looking at literally every page. And for fics with multiple categories, that means opening each fic to see what categories it is tagged to. There's just no practical way to collect that data on a large scale.
(Yes, you can write a script to collect it for you. Yes, AO3 will rate-limit you fairly quickly. AO3 rate-limits me regularly doing regular manual data collection.)
We run into a similar problem when it comes to looking at character popularity. We can see how many fics a character is tagged in, but not only do we have the same kudos popularity issue, but there's also basically no way to tell whether characters are being tagged as a primary vs secondary character purely from the numbers.
Natasha Romanov is the fourth-most tagged Marvel character on AO3, but she is rarely the main character of fics she is tagged in, and those fics where she is the main character tend to be less popular (fewer kudos) than fics where other (male) characters are the main character. So the data we can easily collect say that she is one of the most popular characters in the Marvel fandom, but looking at both kudos numbers and the fics that she is tagged in tell us that that's not really accurate.
And obviously the qualitative part is important--but it's really hard to do that on a large scale or with characters/fandoms that you are less familiar with.