GameRank
UGAHacks 5: Determining how fun professional sports games will be to watch with Deep Machine Learning
What it does
GameRank assesses how exciting future NBA and NFL games will be with artificial intelligence. We specifically focused on two metrics: close games and fan attendance.
ML_Models and Data
We utilized TensorFlow and Keras to build Deep Convolutional Networks for performing Multiple Regression, as well as some other machine learning models that did not perform as well. We programmed the models in Python in Google Colab and we used Google Colab's Cloud Computing for training and testing the networks. Data Sources: Basketball-Reference, Pro-Football-Reference, PushShift.io API, SportsData.io API, Sports Media Watch
BackEnd & FrontEnd for Application
Originally written in Node.js and React, but we switched to building in PyQt5, because we thought it would serve our purposes better.
Built With
- google-colab
- keras
- pyqt5
- python
- sports-data-io-api
- tensorflow
Log in or sign up for Devpost to join the conversation.