This code was built for the NBA 2018 Business Analytics Hackathon.
Problem
Business Analytics: We have provided a sample of 460 games (230 from the 2016-17 season and 230 from the 2017-18 season) for which your task is to predict total international viewers. To assist with your model, we have provided viewership numbers from 1,000 games by country in each of the two seasons, plus stats on team and player performance and status. Using these inputs, we would like you to predict the total number of international viewers for each of the 460 games in the test set.
Files:
data_processing.py - Pre-process the data
prep_test_data.py - Prep test_data
Random_Forest_Reg.py - Random Forest Regression File
output.csv - Final processed dataset
test_data.csv - test data
test_Set.csv - Final output with total viewers for missing games.
Feedback or comments are appreciated.