[wheybags]

Would be nice to see some qualitative analyis to know if it's just slop, or actually more interesting projects. Not sure how to do that though. I think just looking at votes wouldn't work. I would guess more posts causes lower average visibility per post which should cause upvotes to slump naturally regardless of quality.

Edit: maybe you could:

- remove outliers (anything that made the front page)

- normalise vote count by expected time in the first 20 posts of shownew, based on the posting rate at the time

[Normal_gaussian]

a weighted sampling method is probably the best. Segment by time period and vote count or vote rate, then human evaluate. This could be done in a couple of hours and it gives a higher degree of confidence than any automated analysis.

[carimura]

sentiment analysis of comments?