Some Touhou Analysis

Last month, I did some analysis on the entirety of Touhou posts on Danbooru that I wanted to share since it was illuminating to me, and it might provide insight and generate conversation among others as well.

The primary motivator for this was that the EoSD characters have the highest post counts, and it seems like the newer characters could just never catch up. I wanted to try normalizing the data and account for the longevity when looking at the total post counts.

So, I used the number of months since a character was released for normalization. Even though Danbooru wasn't around when the first couple of games were released, the characters in those games have been around, making an impact with artists that have generated artwork. Just as an aside, I didn't include the more recent games since their numbers haven't settled down yet.

One thing to note about Reimu and Marisa is that I used the time since they were released in EoSD as the normalizer, since using the PC-98 dates gave them too much of a penalty, and they were no longer in the top positions anymore, which doesn't make a whole lot of sense for the main characters of the series.

I also included the Coefficent of Variation to show which games had the most variance between the characters.

That's a helluva lot of works and typing.

Not trying to be mean, but this doesn't mean anything. Popularity in general doesn't count for much. The only thing that matters is the official works, the fans' reactions don't mean a lot.

Not to mention this ignores a lot of variables by using only danbooru's databases. Pixiv alone has 1.8 million posts to danbooru's 0.4 mil. Trying to "penalize" older games also seems like cherry picking to me. You put too much weight on the current flow of fanworks while marginalizing the old stuff.

I don't know why you feel the need to do this, but I'll say this: you don't need numbers and stats to justify being a fan of something.

Sacriven said:

...

Just something I did in my spare time.

Hoobajoob said:

...

I agree with all that you said. I wanted to mention that the above analysis is somewhat limited, but I felt I had already typed out quite a bit. It's not meant to be taken seriously, nor is it methodical. There is of course the total number of artwork that has ever been made, which I don't know, and from that, there is the amount of that artwork that has been uploaded to this site, which I do know. Popularity also varies over time, and it's definite that if I took a time slice of last month's posts, I would get completely different results. The above is more like the cumulative popularity of Touhou characters over all time with all of Danbooru based upon the entirety of artwork available.

Also, I probably could have used a better word than "penalize", because I like the EoSD characters, when I really meant that I wanted to show the best normalization and see if time really played a factor. As I originally mentioned, I just wanted to see if the number of posts for EoSD characters was due to them being around longer or if there was something more to it than just longevity. It validated, at least to myself, that that wasn't the case, since even with the largest normalization factor being applied to them, they still did well against the other characters.

Finally, the above wasn't meant to disparage any characters, and it's not meant to say that certain characters are better than others. There are characters I'm a fan of which don't have as much artwork as I'd like. Some characters just do not have as much representation in fanworks as others. That doesn't mean that I'm going to be any less of a fan of them though.