I have performed detailed analysis on stock forum big data (I cannot disclose the source).
My analysis (Poster Count & Avg Post Count Per Poster over SP with posts classified by content based sentiment) concluded that stock forum post volumes for the most part represent anxiety levels.
When the chart is ranging, post volumes are average with around an average count of unique posters.
As the charts fall, post counts increase. (After classifying posts, I concluded the cause being that anxiety increases as portfolios drift to the red and posters activity increases with their anxiety).
The highest post counts correlate extremely well to turning points triggered by events (News drives visibility).
As stocks start to rally, post counts die down amongst high volume posters (I attributed this to easing anxiety) and new poster appearances increase (trend chasers brought in by the chart).
I was unable to find great value in continuing the analysis or using such data as a trade indicator as price action based quant told me basically the same thing with less complexity.
