Percentages were found by comparing the average number of replies containing Naruto per day in each month, and then using that number against the estimated daily reply count on /a/ to find a rough percentage.
Can you overlap this with a graph for troll posts? Troll posts defined as posts containing: topkek, lel, an emoticon, and other obvious baitwords.
That would be interesting to see.
Up to 1/3 of all replies containing the word "Naruto" and that doesn't include the ones that contain the word "Sasuke."
nice fake graph
I don't have anything like that, but I did this one a while ago.
Nigga that's .3 percent, not 30 percent
Does this graph include posts that complain about people using the word doge, incidentally causing them to use the word doge as well?
Yeah it does. But more people complaining about doge means that there's more doge to complain about, so the stats are still good.
One post of doge could result in several posts complaining about the use of doge.
On another board, a post about doge could result in no replies using the word doge.
>you realize you are now inflating doge stats
>/v/, /fit/, and /sp/ have the same amount of doge
>It's twice the amount /a/ has
Where are you getting the data, OP? foolz catalog?
I've been keeping track of /a/ the past 2 weeks, though just aggregate data. Ignore the spikes on 02/10, that was just a bug in my software.
I've been using foolz mostly.
put /g/ on that and it'd shit itself.
Cool, are you just working through their search system, or do you have access to their database?
Here's the absolute number of posts on /a/, plus the percentage of posts with images at any given time.
Nah, I've just been plugging in keywords and gathering up the results to use as data. I don't have access to the database or anything like that.
Why on earth didn't you include /g/
it would have gone to the moon
/g/ had like a billion fucking instances of doge over two months, it would have needed it's own graph.
that's the joke.mkv
Ok. I wonder if foolz ever offers sql dumps. I've been wanting to do the same sort of data mining for a while now.
Here's the absolute size of the full images on /a/, plus stats on the average size of new images.