Right now spike/potential/hyper classifications are my subjective opinion.. which I hate because that varies depending on the day I'm having and my specific biases.

Figuring out how to classify them mathematically is next on the to-do list. For example a spike would have extreme outliers in the middle and hyper would be over a certain exponent.

It would be nice if it could discriminate between trends that are still in exponential growth from the ones that have passed the inflexion point.