Skip to content
Amateur squash match prediction for Ottawa squash players
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data
models
visualization
README.md
process.py
requirements.txt
train.py
visualize.py

README.md

Ottawa Squash Match Predictions

I created a machine learning web application which can be used to predict the score of amateur squash matches in Ottawa. This repository contains all the code I used to generate the model and clean the data. See the website here. See the full source code here. Squash players in Ottawa use a website called Rankenstein which has kept track of squash matches since 2005. I trained my model using 15,000 games played in Ottawa with wide variety of stats from 1000+ players.

Data Visualization

High dimensional data of wins/losses visualization by TSNE:

TSNE visualization of wins and losses

Features

SHapley Additive exPlanations (SHAP) summary plot visualizes feature importance

Feature Importance I've tried many features including:

  • Rating
  • Win streak
  • Win rate vs. opponent
  • Trend (increase in rating over time)
  • Upset rate
  • Time since last played
You can’t perform that action at this time.