Hockey match win prediction using historical data analysis. Apply only if you are strong with analysis

I would like to predict daily hockey outcomes for each daily matchups. I have each teams statistics, odds and league schedules for the last 5 years for historical analysis. Each of the team statistics, odds and schedules are on 3 separate csv files. Each will be updated daily. Each will need to be read into the program. I would prefer R but if in Python I would need a tutorial as I am not familiar with running statistical programs in Python. On a daily basis, the odds would need to be used and updated from that day. The statistics will be of course used from the previous games as todays game will not have any statistical outcome. Hockey is statistically based sport with some trends and situations that I can share at a later time we can program in. I would prefer to use some deep learning AI to enhance the models accuracy greatly. Regarding the output of the model I am looking for an easy to read layout that is neat. Also regarding the stream piping and preprocessing of the data that must all be included with the model and project. The model success itself, I want to be the best it can be before the daily matches. After the games and stats and odds are updated it would be great if the model would continue to learn from itself. However I don't want a model that will give me false results, for example telling me it will predict 100% of the time. Of course it will, after it knows the outcome! Perhaps the model will be that good, but I am looking for a developer to be straightforward and get the best model before the games uses the data and knowledge.