I see several problems: the biggest is that the project is based on CMC. The data available there has several known problems, so garbage in, garbage out!

Take the recents jumps to 8kUSD of bitcoin for example, based on the ETF stories. If you followed the market at that time, you would know that the CMC represenation of it is at best flawed. Also there are plausible accusations of faked volume and market manipulation.

The second is the project does not monitor the blockchain itself, or sentiment analysis of news feeds, so it is missing alternative stories.

Finally, the architecture is way too complex. The author mentionned in the first post doing things the right way, but I see that mostly as an exercice to try to use Akka for a problem where it is a bad fit.

The work is interesting, but I fear it is the wrong approach