Savadogo / Sparkify-project

Build a machine learning model on IBM watson with spark

Predicting churn with Spark

This project has been realised as the capstone of my pathway to being a datascientist with Udacity nanodegree. I had to build a machine learning model to predict user churn for a music streaming platform (Sparkify).

Table of Contents

1. Packages requirements
2. File Descriptions
3. Licensing, Authors, and Acknowledgements

Packages requirements

• Python 3.*
• pyspark
• matplotlib
• numpy
• pandas

File Descriptions

The following files are in this repository:

• Sparkify - worspace.ipynb - a notebook of the analysis on the small dataset of 128 MB. This notebook is exported into Sparkify - worspace.html.

• Sparkify - ibm watson.ipynb - a notebook of the analysis on the meduim dataset of 231 MB realised on IBM Watson. This notebook is exported into Sparkify - ibm watson.html.

Results

A detailed article on the finding blog post available here.

Licensing, Authors, Acknowledgements

The code in this project is licensed under MIT license.