dsam7 / cse6242-nba Goto Github PK

User Guide for CSE6242 Team 15's Final Project

Description:
This is a project containing multiple analyses of NBA data to highlight some interesting trends observed in the unique 2020 NBA "Bubble" season. Our approach to this project is broken down into 3 sets of visualizations:

    - Shot Maps : A set of shot maps comparing shooting positions and accuracy of players and teams as a whole, before and inside the Bubble.
    - Prediction Model Comparison : Python notebooks and an interactive visualization to compare predictions of Machine Learning models on past games, 2019-20 non-bubble games and Bubble games.
    - Choropleth: *explanation*

Installation:
All our code is attached to the official canvas submission under the folder CODE. The code can also be cloned from Github using the following command:
        git clone https://github.com/dsam7/cse6242-NBA.git

Execution:

1. Shot Maps:
    i. Python notebook:
        a. Once you have cloned this repository, open the Shot Maps folder and open ShotMaps.ipynb.
        b. All necessary instructions are laid out in the Notebook.
        c. All NBA team ids can be found here: https://github.com/bttmly/nba/blob/master/data/teams.json.
        d. Run through the cells sequentially to generate CSV data that can be used to create ShotMaps that can be displayed in visualization software like Tableau.
        e. The imgs folder currently stores the visualizations generated on the Phoenix Suns that are shown in the Python Notebook.
    ii. Tableau Book
        a. An example Tableau Book has been provided.
        b. The graphics in it use data from the games the Phoenix Suns played in the Bubble. (one of the CSV files generated in the Python notebook)
        c. Replace the current data source with NBA API data of your own to make your own visualizations.

2. Prediction Model Comparison:
    i. Python notebooks for pre-processing and models:
        a. Navigate to the "Prediction Models" folder and run `jupyter notebook`.
        b. In preprocessing.ipynb, run the cells sequentially to see the datasets used and calculated advanced stats.
        c. NOTE: Only run the prepare_game_data() function and further cells if the file `Prediction Models\generated_data\games_formatted.csv` does not exist.
        d. In models_predictions.ipynb, run the cells sequentially to create the models and write the predictions to the output file.
    ii. d3 visualizations:
        a. Navigate to the "Prediction Models" folder and run `python -m http.server 8080`.
        b. Navigate to http://localhost:8080/ and in the "Web Pages" folder and open `predictions.html`.
        c. On the page, use the dropdown menu to select a particular game. The UI should update to display the selected game, the actual result of the game, and the winners predicted by the ML models.
        d. In the `analysis.html` web page, view interesting tabulated results and inferences drawn from the predictions of the ML models.

3. Choropleth: *Insert demo instructions*
    i. Python notebook:
        a. Data is already preprocessed and in the folder, but to process the data again you can navigate to the "Shooting Choropleth Map" folder, open the nba_against_stats.ipynb, and run all the cells
        b. This will store the preprocessed data in the .csv files
    ii. d3 visualizations:
        a. Navigate to the "Shooting Choropleth Map" folder and run `python -m http.server 8080`.
        b. Once in the http://localhost:8080/, navigate to choropleth.html
        c. On the page, use the dropdown menu to select a particular statistic. The map will update to display the statistic you have chosen.