Some years ago at the retail company I worked at then, our data analyst came up to me. She was working on forecasting how much each of our products would sell in the next 12 months and wanted to know if I could help develop a piece of SQL to do this.
Such forecasting can be done with a multitude of different models, each suitable to different types of data and circumstances. She had experimented with tools and researched and ran tests of the models on selected products and done whatever magic analysts do to make discoveries in our data. In the course of this, she found ...
