Predict Customer Churn

• Project Predict Customer Churn of ML DevOps Engineer Nanodegree Udacity

Project Description

The purpose of this project is to predict customer churn rate for credit card holder.

Project Setup

• Install poetry
• Install requirements with either:
  • poetry install
  • export to requirements.txt with poetry export --without-hashes -f requirements.txt -o requirements.txt and pip install -r requirements.txt

Docker image

Files and data description

The data set, for convenience, is saved in data directory All the outputs will be written to outputs directory, unless differently specified with command line argument -o | --output

Running Files

Python

NB if running using poetry prepend poetry run to any python command

python churn_library.py data/bank_data.csv (-o /path/to/outputs)

Docker

Build

docker build . -t churn

Run

docker run -v /path/to/output:/app/churn/outputs -v $(pwd)/data:/data churn /data/bank_data.csv WARNING docker user is root

Run tests

NB if running using poetry prepend poetry run to any python command

pytest

Test logs will be written to ./outputs/logs/