Agreed. Started out so promising and then never updated. It does seem to make sense to be honest, trading based on prediction of psychology, but it depends a lot on the amount of data you have I would say also. So much going on if your sample is too small then you miss out on the really important elements.
The sample size certainly matters, but I think more important than the quantity of data being gathered (and in the age of big data it better be huge) is the quality of that data. Whale data should exclude false positives (an exchange shouldn't be considered a whale, for example), and psychology data should be able to differentiate between bot accounts (Twitter's just full of spam bots posting fake sentiment that would affect psych data, for instance).
Nice idea but ultimately, hasn't worked even on enterprise scale. I'm guessing that's why OP never updated with proof.