NRL Machine Learning Models, Data Analytics and Data Scraper

This library is still a Work-In-Progress. Feel free to help out by adding to the repository.

Description

This project is a webscraper for NRL data, and provides a TensorFlow machine learning model for NRL related predictions.

How to Use

To add this section in later

Data

All data for this project is hosted on this website. I personally host this website with all data being stored in a S3 instance.

Player Statistics

Match Data

Match Data JSON Schema

{
  "$schema": "http://json-schema.org/draft-04/schema#",
  "type": "object",
  "properties": {
    "NRL": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "2024": {
            "type": "array",
            "items": {
              "type": "object",
              "properties": {
                "1": {
                  "type": "array",
                  "items": {
                    "type": "object",
                    "properties": {
                      "Details": {
                        "type": "string"
                      },
                      "Date": {
                        "type": "string"
                      },
                      "Home": {
                        "type": "string"
                      },
                      "Home_Score": {
                        "type": "string"
                      },
                      "Away": {
                        "type": "string"
                      },
                      "Away_Score": {
                        "type": "string"
                      },
                      "Venue": {
                        "type": "string"
                      }
                    },
                    "required": [
                      "Details",
                      "Date",
                      "Home",
                      "Home_Score",
                      "Away",
                      "Away_Score",
                      "Venue"
                    ]
                  }
                }
              }
            }
          }
        }
      }
    }
  }
}

Detailed Match Data

Workspace

This code is updated on Jupyter Notebooks (ipynb) and default python (py) files.

Table of Contents

• NRL Machine Learning Models, Data Analytics and Data Scraper
  • Description
  • How to Use
  • Data
    • Player Statistics
    • Match Data
    • Detailed Match Data
  • Workspace
  • Table of Contents
  • Web Scraping
  • Machine Learning
  • Web Scraping
  • Visualisations
  • Conveters
  • Installation
  • Future Tasks
  • Help

Web Scraping

This project utilizes Selenium for web scraping NRL data from the NRL website. Currently, I manually perform this task weekly with limited plans for automation at present. This code is located in: /scraping/

There are four different web scrapers:

• Match data 2024 - match data for every game in 2024. Updated regularly
• Match data 2015-2024 - match data for every game from the select years. Data is stored on the above website.
• Player data 2024 - player data for every game in 2024. Updated regularly
• Player data 2015-2024 - *player data for every game from the select years. Data is stored on the above website. * NOTE: To obtain player data you need match data first.

Machine Learning

This code is located in /predictions/ There are two different machine learning models:

• Match based: Uses match statistics to form the final result. This project requires furthur optimisation.
• Player and Match based: Uses player statistics to form the final result. This project is currently WIP (however it provides code on how to manipulate the player data).

Web Scraping

This project utilizes Selenium for web scraping NRL data from the NRL website. Currently, I manually perform this task weekly with limited plans for automation at present. This code is located in: /scraping/

Visualisations

Ways to Display the data are located in: /visualisations/

Conveters

JSON is the default format for all code. Conversions tools have been generated to assist those who need .txt or .csv formats. These are located in: /converters/

Installation

• Download the required data from the above website and place it into the /data/ folder.
• Install the requirements.txt file
• Run the Jupyter notebook located in /predictions/

Future Tasks

• Update the machine learning model to work with 2024 data
• Update the website to display prediction results
• Clean up all the code
• Optimise the current machine learning model
• Update requirements.txt
• Provide a more descriptive README
• NRLW data
• Anytime Try Scorer Probability model
• Try Location Data
• Team Stats - All Runs, All Run Metres, Post Contact Metres, Line Breaks, Tackle Breaks, Average Set Distance, Kick Return Metres, Average Play the Ball Speed, Offloads, Receipts, Total Passes, Dummy Passes, Kicks, Kicking Metres, Forced Drop Outs, Kick Defusal, Bombs, Grubbers, Effective Tackle, Tackles Made, Missed Tackles, Intercepts, Ineffective Tackles, Errors, Penalities Conceded, Ruck Infringements, On Reports, Interchanges Used
• Replicate https://wicky.ai/content/analytics/predictive-analytics-applied-to-rugby-league-looking-at-try-scorers-in-the-nrl/
• Provide a text export

Help

I intend for this project to be open source, so help is always handy!