This repository contains Python scripts that demonstrate how to perform Gaussian Mixture Model (GMM) clustering using Polars for data manipulation and scikit-learn for the machine learning component.

1. gmm_with_polars.py - Comprehensive example with visualization and performance comparison
2. simple_gmm_polars.py - Simple, focused example of GMM with Polars
3. requirements.txt - Required Python packages

pip install -r requirements.txt

Or install packages individually:

pip install polars numpy matplotlib seaborn scikit-learn pandas

python simple_gmm_polars.py

python gmm_with_polars.py

• Creates synthetic data with 3 natural clusters
• Preprocesses data using Polars operations
• Performs GMM clustering with scikit-learn
• Analyzes results using Polars aggregations
• Saves results to CSV

• All features of the simple example, plus:
• Data visualization with matplotlib
• Performance comparison between Polars and Pandas
• Detailed clustering analysis and metrics
• Model parameter inspection

• pl.DataFrame() - Creating DataFrames
• with_columns() - Adding computed columns
• group_by().agg() - Aggregating data
• select() - Selecting columns
• to_numpy() - Converting to NumPy arrays
• write_csv() - Saving results

• Automatic cluster detection
• Probability estimates for each prediction
• Multiple covariance types
• Model convergence information
• Cluster centers and parameters

The scripts will generate:

1. Console output with clustering analysis
2. CSV files with results
3. Visualizations (comprehensive example)
4. Performance comparisons

You can modify these scripts to:

• Use your own data (replace the create_sample_dataset() function)
• Change the number of clusters (n_components parameter)
• Use different features for clustering
• Adjust preprocessing steps
• Change visualization styles

Polars offers several advantages for data preprocessing in ML workflows:

• Speed: Faster than Pandas for many operations
• Memory efficiency: Better memory usage
• Lazy evaluation: Optimized query planning
• Modern API: Clean, consistent syntax
• Type safety: Better type handling

This makes it an excellent choice for data preparation before applying machine learning algorithms like GMM.