Identifying cryptrocurrency spam/hype cycles on Twitter

September 8, 2018
python cryptocurrency

Inspired by a very interesting Defcon presentation about stock spammers from 2011, I thought it would be interesting (and maybe profitable?) to monitor the frequency of low-cap cryptocurrency mentions on Twitter, the theory being that pump groups and spammers are likely to ‘hype’ coins before/during large buys in order to manipulate more people into buying.

The free tier on the Twitter streaming API allows you to monitor 400 different terms so I chose to observe the Twitter stock symbol ‘$’ followed by the coin abbreviations, for example $SIA is Siacoin. Ideally we could monitor more intelligently by using smarter rules around which coin a tweet is referring to, but as a starting point this seemed pretty good as the ‘$’ symbol seems to have been fairly widely adopted across Twitter for crypto market discussion.

I collected a few samples across a few days at different points but didn’t see any correlation between sudden Tweet frequency spikes and price. This is just a first exploration and I think it would be useful to plot alongside the coin price on the same graph rather than doing manual comparisons.