NCAAMarchMadnessNN

My supporting code for Google Cloud & NCAA® ML Competition 2019-Men's (4th place finish)

Below you can find a outline of how to reproduce my solution for the NCAA® ML Competition 2019-Men's competition. If you run into any trouble with the setup/code or have any questions please contact me at dave.a.lorenz@gmail.com

#HARDWARE (this should have no problem running locally on an average laptop or deskop) Processor Intel(R) Core(TM) i5-6200U CPU @ 2.30GHz, 2400 Mhz, 2 Core(s), 4 Logical Processor(s) 8GB RAM Windows versoin 10

#SOFTWARE (python packages are detailed separately in requirements.txt): Python 3.7 64-Bit (https://www.anaconda.com/download/)

#DATA

1. Point spreads from: http://www.thepredictiontracker.com/ncaaresults.php
2. Ken Pom data from: https://kenpom.com/index.php (note: 2019 data collected prior to tournament)
3. Kaggle data from: https://www.kaggle.com/c/mens-machine-learning-competition-2019/data

#SCRIPTS (ran in Jupyter notebook but also posted .py and .html versions)

1. 01 Prepare point spread data.ipynb
2. 02 Prepare Ken Pom data.ipynb
3. 03 Combine files for model training and testing.ipynb
4. 04 Create data for 2019 prediction.ipynb
5. 05 Model.ipynb

#How to reproduce competition results:

--Download input files from https://www.kaggle.com/c/mens-machine-learning-competition-2019/data: Stage2DataFiles.zip, MasseyOrdinals_thru_2019_day_128.zip

--Move following files above to 'Data/Kaggle NCAA': MasseyOrdinals_thru_2019_day_128.csv, NCAATourneyCompactResults, NCAATourneyCompactResults.csv, NCAATourneySeeds.csv, RegularSeasonDetailedResults.csv, TeamSpellings.csv

--Run scripts 01-06 above