/march-madness-machine-learning

Predicting March Madness brackets with machine learning :basketball:

Primary LanguageR

March Madness Machine Learning

Final project for INDENG 242: Applications in Data Analysis that predicts the final NCAA March Madness Tournament bracket. We use a logistic regression model to predict the probability that each team wins a given match-up, for all possible match-ups. Then we feed these probabilities into a simulation to predict the most likely bracket.

Project Overview

.
├── data
│   ├── FinalStats.csv
│   ├── NabeelStats.csv
│   ├── RegularSeasonDetailedResults.csv
│   ├── SampleSubmission.csv
│   ├── TeamRatings.csv
│   ├── Teams.csv
│   ├── TourneyCompactResults.csv
│   └── TourneySeeds.csv
├── descriptiveAnalysis.Rmd # Descriptive analysis R Markdown file
├── descriptiveAnalysis.pdf # PDF generated by descriptiveAnalysis.Rmd
├── project.R               # Data processing
├── sampleSimulation.xlsx   # Simulation in Excel
├── targetVariables.R       # Predictive Models
└── teamRatings.R           # Pull team ratings from external source

The CSVs from the data folder are from Kaggle with our own modifications.

A note on using R Markdown for descriptive analysis

For recent versions of RStudio, the Knit PDF button does not seem to work. Instead, enter the following command in the Console to generate a new PDF:

render('descriptiveAnalysis.Rmd')

Authors

  • Farshad Miraftab (@farshadl123)
  • Matthew Rosendin (@rosendin)
  • Nabeel Saleem (@darealnabeel)