gregoryg / ds-for-telco

ds-for-telco

Created by Juliet Houghland + Sandy Ryza ([e-mail])
The source notebook demonstrates building a churn prediction model using Spark and Spark MlLib's pipeline API for cross validation and model tuning. The Pipeline API is available in PySpark in version 1.6 or higher.

Status: Demo Ready
Use Case: Telco Churn Prediction

Steps:

1. Open a terminal and run setup.sh
   
2. Create a Python Session and run setup.py
   
3. In your python session run ds-for-telco.py
   
4. When finished, run cleanup.sh in the terminal
   

Recommended Session Sizes: 2 CPU, 4 GB RAM

Estimated Runtime:
ds-for-telco.py --> approx 1 min

Recommended Jobs/Pipeline:
None

Demo Script
TBD

Related Content:
http://blog.cloudera.com/blog/2016/02/how-to-predict-telco-churn-with-apache-spark-mllib/