This page describes some of the details on the work that went into this site. Mainly it was a multi-year effort with a lot of research and programming involved. Here are the experiences I had over the course of the year.
I enjoyed very much learning new languages (Matlab, Tcl, R, Python) and methods (many) to scrape web data, archive and cleanse datasets, extract features for past performance, jockey, trainer and owner statistics, build multiple predictive models using advanced machine learning techniques with state-of-art data science tools. The aggregation, ranking and reporting scheme runs daily on a cron job in a Xeon-server as well as data collection and aggregation. All is very much automatic on my basement server.
The predictive models for each race is based on a large number of similar races occurred previously in the historical data, and the multiple methods are aggregated in a bagging-style report before published. The entire process runs automatically without any interactions and I try to monitor daily to make sure everything went fine. Thanks to wordpress/tumble/facebook and twitter to allow me to reach many of the followers. This was a great experience and I try to continue as long as I can.
Later I will try to give more details. But for now that is what I could say. Best of luck to you!