NBA

This repository is related to the NBA topic and Machine Learning.

The Repository Structure

The repository includes the following .ipynb files:

01_problem_definition.ipynb

This notebook outlines the primary goals and objectives of the analysis.

02_data_collection.ipynb

This file is dedicated to the data collection process. It includes steps to gather, clean, and prepare NBA data from various sources for analysis.

03_data_exploration_p1.ipynb

The first part of data exploration is dedicated to the exploration of the NBA SQLite database.

04_data_exploration_p2.ipynb

The second part of data exploration is dedicated to the exploration of the NBA PostgreSQL database, collected over API.

05_data_exploration_p3.ipynb

The third part of data exploration is dedicated to the identification of patterns, trends, and statistical analysis.

06_linear_and_polynomial_regression.ipynb

This notebook focuses on implementing and evaluating linear and polynomial regression models, exploring relationships between various features in the NBA data.

07_tree_algos.ipynb

This notebook focuses on implementing and evaluating models like the decision tree, random forest, gradient boosted trees, and XGBoost.

08_nn_algos.ipynb

This notebook focuses on neural network models, applying techniques like various architectures, dropout, regularization, learning rate reduction, and early stopping.

09_clustering.ipynb

This notebook focuses on clustering with two or more features, employing the elbow method and K-means.

10_anomaly_detection.ipynb

This notebook is dedicated to anomaly detection with two or more features, using the Elliptic Envelope and Gaussian Mixture Model.

Additionally, there are a few folders:

• The figures folder contains images for some .ipynb files.
• The reports folder holds data profiling reports that can be opened in any browser.
• The sql_scripts folder includes several useful SQL queries related to the PostgreSQL Database.
• The utils folder contains several Python files that are excluded from .ipynb files to avoid overloading them with code. Links to these files are included in the .ipynb files.