>mfw I have no face
You (as sophisticated gentleman who browse 4chan erryday) may wonder: how many posters actually post image of their face with riveting greentext story ending with >mfw and how many of them have no face. This question is completely useless and serves no purpose to humanity, but nonetheless it's marginally interesting and as such should be answered by the power of science.
And guess what? I have massive 4chan archive and now I have also the answer.
Yes, this graphical clusterfuck is a result of my work. It's plot of ratio of posts containing >mfw to posts with image.
As you can see, maximal ratio is around 90% (mainly „graphical“ boards: /co/, /vg/, /mlp/, /a/). Most boards are between 70% and 90%. /sci/ and /g/ are fluctuating around 65%. /r9k/ is bellow 50% (Do you even trying, robots?) /q/ is so low because I have not enough data to compute proper ratio. (This dataset ends on 2012-08-09, few days after /q/'s inception. I have the rest of data in huge pile of scraped html files and JSONs from 4chan API, that waits for parsing and processing.)
Chart is heavily smoothed (values are averaged with previous and next days), otherwise it would look even worse. Some boards are missing. It's because they are too slow and there is not enough data to compute accurate has face ratio.
And there is table of average has face ratio divided by board:
- Libreoffice is slow as fuck (complicated chart takes ages to draw).
- MySQL is very bad at handling rectangular data (boards ⨯ days), but with few smart hacks one can make it fly.
- My english sucks and I know it.
- There's the data make better chart.