Well, in this post I made reference to two of my models: (1) a recession prediction model and (2) a volatility model. The recession prediction model is based on over 100 years of data. Most of it is monthly macroeconomic, although not quite all of it. It also has some market data and some other data that is neither macroeconomic nor market. The volatility model is a real mixture of multiple types of data. It involves about 30 years of data, most of which is daily, but some of which is one-minute.
I use many methods for modelling: regression, machine learning and others. I don't restrict myself and basically use whatever techniques seem most appropriate for the task at hand. In the case of these models, the recession model is basically a logistic regression model, although I adapted it a little bit for my purposes. I have a portion of my portfolio in "traditional investing" but prefer to avoid recessions to boost long term profits. The volatility model is a completely non-standard form of modelling. I developed the underlying mathematics for the modelling process myself. I have two degrees in mathematics, one of them being an advanced degree, so I know how to do this, although it takes quite a bit of time. I've been trading volatility futures and options, both long and short, for about 8 years now. It's one of the areas I've really taken a liking to modeling. I find it extremely interesting.
Thanks to everyone who has provided constructive replies so far, by the way. I am finding this interesting and hope some others are finding it interesting as well.