NBA All Star Predictor

Project Description

• A personal project to explore scraping and basic machine learning.
• Big Question: Which players were all-stars?
• Assumption: taking full season stats instead of stats up till all-star break.

App Deployment

Link to App: NBA All Star Predictor

• Hosted on Streamlit Share

Data Source

• NBA Player Season Totals taken off Basketball-Reference.com
• Scraped Data Includes:
  • Player List from 1996 - Present
  • Individual season stats totals
  • Individual season per game stats

Sample Data

Player List

Season Totals

Per Game

Modelling

Models Compared

LightGBM Classifier
Random Forest Classifier
XGBoost Classifier

Deployment

• Deployed on Streamlit
• Use sidebar to input player's stats
• Predictions presented from all three classifiers

Coding Language & Main Libraries

• Python Version 3.8.3
• Pandas Version 1.2.2
• Scikit-learn 0.23.2
• Streamlit 0.78.0