One thing that is striking to see is that time did not make things better for bitcoincash: there is some creative destruction in the form of services that bitcoincash came to dominate later, but there are not really many. And few more that did not exist at the beginning came online later as developers started building tools specifically for this economy. But in general, the ones that were strong at the beginning remained so across the whole period of time.
Predictive model. The first thing that we should note is that these are fat arrays — that is, they have too many variables and very few observations. This is by itself a challenge for most machine learning techniques unless an evolutionary algorithm is used.
When we start learning ranking models, we interpret returns as an indication of value, and the demand signals from traffic sources (e.g. average page visits) as the drivers of that value. Since we are constructing a ranking system, we are interested in the demand relationships that optimize BTC_returns > BCH_returns.