Дата публикации: 28-07-2026 03:01:38

Hey! Thanks for the question -- it prompted me to start collecting data about comments (after procrastinating on it for a while, because I had to write new code to gather comment data). I've also seen other discussions from folks also thinking about how to do this kind of analysis (like in the fandom data projects community) -- hopefully we'll end up with multiple people attacking this from different angles and getting a variety of data about comments!I'll give a sneak preview that partially addresses your question and contains some good news. If we look at the fraction of AO3 works that get at least one comment (focusing just on one-shots for now), I think things have gotten better over the past decade on AO3*:In other words, it tentatively looks like more works were getting at least one comment in 2024 than in 2014 (for a variety of time periods). One caveat, though -- if a bunch of works with no comments got deleted in the interim, there will be survivor bias here. I'll try to look into that possibility later. Another caveat: this is based on only like ~100 randomly selected works from each year -- this may all change with more data!Another interesting tidbit: I still see some of the 2014 works getting comments. In fact, ~30% of works have gotten new comments over 5 years after they were posted, and it looks like ~10% of one-shots posted back in Mar 2014 got a new comment in 10 years later, in 2024.I'm still doing other analyses; there may be other factors that better match with the discourse around how comment culture has changed. It could be that comment activity peters out faster now than it did back then, for instance. Or the total number of comments left on the popular works is less now than it was back then (though my current methods may not be able to capture that). Edit thanks to quick eagle-eyed readers: it's likely that some of what people are thinking about is ratio of comments to hits -- that is hard to compare in 2014 to 2024, because we don't know which hits came from which years. But I am working on some analyses along those lines. :)If you have other hypotheses about what's changed in commenting culture, feel free to share! I'll look into what I can.Some methodology notes:*I've been tackling this by comparing AO3 one-shots posted in early 2014 to one-shots posted in 2024, and comparing activity in the days/weeks/months immediately after the works were posted. (To start with, I'm only scraping the first page of comments for each work -- meaning the first 20 comment threads -- so there are lots of comments I'm potentially missing for the really popular works. But for many works, this captures all the comments, and I think it may be sufficient for a lot of the analyses I am interested in.) I'm choosing to focus on 2014 vs. 2024 because 2024 is close to now (but it's been long enough for comments to have settled down a bit), and 2014 was well after AO3 was established (thus it was already a pretty lively time on AO3). I don't want to collect data about every single year because it's too time intensive/too hard on AO3's servers. But if people think that I should be looking at different years, I'm interested in feedback.Because it's only been ~10 months since March 2024, I am limiting a lot of my analyses to only look at commenting activity the first ~10 months after works were posted in both cases.destinationtoast:inimitablereel:I pulled some numbers in November shortly after making that fandom data projects post based on filtering for numbers of works over a year and numbers of comments (sorry I did make the opposite graph from destinationtoast so lower is better here) - the 2024 number is going to include works that are very new so don't put too much trust in it. In general I'm sharing these numbers since I already have them but it doesn't account for the things toastystats was measuring: fic gets more comments over time, so grabbing a snapshot of total comments NOW is ignoring that for fic from a decade ago people have had 10 years to commentALTI also pulled median numbers of comments per fic (which you can figure out by grabbing the total number of fics per year which I already had and then filtering for works with

Основное содержимое страницы с новостью.

Anonymous asked:

Hello! I keep hearing that fandom culture has changed, and there are less comments now than there were years ago. Have you noticed this in your analysis? Is the percentage of comments being left today lower than before?

destinationtoast answered:

Hey! Thanks for the question – it prompted me to start collecting data about comments (after procrastinating on it for a while, because I had to write new code to gather comment data). I’ve also seen other discussions from folks also thinking about how to do this kind of analysis (like in the fandom data projects community) – hopefully we’ll end up with multiple people attacking this from different angles and getting a variety of data about comments!

I’ll give a sneak preview that partially addresses your question and contains some good news. If we look at the fraction of AO3 works that get at least one comment (focusing just on one-shots for now), I think things have gotten better over the past decade on AO3*:

In other words, it tentatively looks like more works were getting at least one comment in 2024 than in 2014 (for a variety of time periods). One caveat, though – if a bunch of works with no comments got deleted in the interim, there will be survivor bias here. I’ll try to look into that possibility later. Another caveat: this is based on only like ~100 randomly selected works from each year – this may all change with more data!

Another interesting tidbit: I still see some of the 2014 works getting comments. In fact, ~30% of works have gotten new comments over 5 years after they were posted, and it looks like ~10% of one-shots posted back in Mar 2014 got a new comment in 10 years later, in 2024.

I’m still doing other analyses; there may be other factors that better match with the discourse around how comment culture has changed. It could be that comment activity peters out faster now than it did back then, for instance. Or the total number of comments left on the popular works is less now than it was back then (though my current methods may not be able to capture that). Edit thanks to quick eagle-eyed readers: it’s likely that some of what people are thinking about is ratio of comments to hits – that is hard to compare in 2014 to 2024, because we don’t know which hits came from which years. But I am working on some analyses along those lines. :)

If you have other hypotheses about what’s changed in commenting culture, feel free to share! I’ll look into what I can.

Some methodology notes:

*I’ve been tackling this by comparing AO3 one-shots posted in early 2014 to one-shots posted in 2024, and comparing activity in the days/weeks/months immediately after the works were posted. (To start with, I’m only scraping the first page of comments for each work – meaning the first 20 comment threads – so there are lots of comments I’m potentially missing for the really popular works. But for many works, this captures all the comments, and I think it may be sufficient for a lot of the analyses I am interested in.)

I’m choosing to focus on 2014 vs. 2024 because 2024 is close to now (but it’s been long enough for comments to have settled down a bit), and 2014 was well after AO3 was established (thus it was already a pretty lively time on AO3). I don’t want to collect data about every single year because it’s too time intensive/too hard on AO3’s servers. But if people think that I should be looking at different years, I’m interested in feedback.

Because it’s only been ~10 months since March 2024, I am limiting a lot of my analyses to only look at commenting activity the first ~10 months after works were posted in both cases.

I pulled some numbers in November shortly after making that fandom data projects post based on filtering for numbers of works over a year and numbers of comments (sorry I did make the opposite graph from destinationtoast so lower is better here) - the 2024 number is going to include works that are very new so don't put too much trust in it. In general I'm sharing these numbers since I already have them but it doesn't account for the things toastystats was measuring: fic gets more comments over time, so grabbing a snapshot of total comments NOW is ignoring that for fic from a decade ago people have had 10 years to comment

ALT

I also pulled median numbers of comments per fic (which you can figure out by grabbing the total number of fics per year which I already had and then filtering for works with <x comments until it's about half) - the fact that this peaks in 2017 like the other graph is expected, but I think it's useful for the discussion to note how low ALL these numbers are

ALT

Some notes on methodology: I used ao3's search page and did a bunch of filtering using the "Date" field for "x to x+1 years ago" which means that my years are all actually slightly off: for example 2024 as listed here is Nov 19 2023-Nov 19 2024. I think the Date field is also date updated not date posted, to the extent that that makes a difference with oneshots? Happy to share my spreadsheet but it would probably be a better idea to redo it with a search for "created_at" (and figure out a way to automate if you're grabbing a bunch of numbers)

I also grabbed comments on the 21st most commented work (21 because it's the top of the second page of results, the 1st varied wildly and let me find a lot of people who were just having whole long running conversations in the comments, which is super interesting but doesn't really tell you anything about comment culture) - not posting that graph because it looks coherent until you divide by hits and then it's incoherent and in general I think any individual fic is not a great source of stats.

Highly relevant to my recent post, but in typical me fashion, I had completely forgotten about this! 😂😭 @inimitablereel makes a wise point about BNF experience vs. typical experience and methodology limitations:

ALT

See more posts like this on Tumblr

#(and hilarious) #inimitablereel #and for your past work! #other people's analysis

Схожие новости