Predict Customer Churn

• Project Predict Customer Churn in ML DevOps Engineer Nanodegree program by Udacity.

Project Description

The project is on credit card customers that are most likely to churn, based on the dataset available in Kaggle. This project implements two different churn predictors: a Random Forest Classifier and a Logistic Regression method.

The aim of this project is to demonstrate my competencies in writing clean code, to be specific:

• The ability to write modular and efficient code, with proper documentation and style check (using pylint and autopep8)
• Implement best practices such as errors handling, testing and logging

Files and data description

The directory structure:

.
├── data
│   └── bank_data.csv
├── images
├── logs
│   └── churn_library.log
├── models
│   ├── logistic_model.pkl
│   └── rfc_model.pkl
├── churn_notebook.ipynb
├── churn_library.py
├── churn_script.py
├── churn_script_logging_and_tests.py
├── README.md
└── requirements.txt

The original given file (churn_notebook.ipynb) is refactored into .py files for modularity, reusability, as well as ease of testing and logging.

More description on some files and folders included in the project:

• churn_library.py: contains utility functions that help with data analysis and model training
• churn_script.py: contains script that runs the functions from churn_library.py
• churn_script_logging_and_tests.py: used for testing functions in churn_library module
• data: the directory that contains the data file used in this project
• images: used for saving images of data analysis and model analysis
• logs: contains log generated from running testing script
• models: used for storing model that are ready to use in production

Running

Dependencies

List of libraries used for this project:

autopep8==2.0.1
joblib==1.2.0
matplotlib==3.6.2
numpy==1.23.5
pandas==1.5.2
pylint==2.15.9
scikit-learn==1.2.0
seaborn==0.12.2
shap==0.41.0

Use the package manager pip to install the dependencies from the requirements.txt

pip install -r requirements.txt

Modeling

Run the following command to execute the main script

python churn_script.py

script execution generates

• EDA and model metrics plots are available in directory ./images/
• Model pickle files are available in directory ./models/
• Log files are available in directory ./logs/churn_library.log

Testing and Logging

Run the following command to run the tests script

python churn_script_logging_and_tests.py

script execution generates

• Log file ./logs/churn_library.log

License

Distributed under the MIT License. See LICENSE for more information.