1 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
Venkatesh Sellappa (Venky)
Solutions Engineer
2 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
The Quant Workbench
3 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
Matlab
SAS
Java/C++Python
R
4 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
Excel Is King
5 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
Trial And Error
Data Ingestion
6 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
The Project
8 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
Security Monitoring
9 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
The Solution ?DS X Workbench
10 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
Use Case
All industries are effected by churn.
Being able to predict churn helps companies take action and keep customers longer.
The more historical data, the better the model
Data collected and labeled over time based on churn.
Using a Random Forest we will predict future churners.
Customer Churn Architecture
11 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
Demo ScenarioAssessing Customer Churn Probability in Real Time
Stored long term data on customer churn behavior
New real time data coming in
Predict a customers churn probability before they churn
Alert the proper departments | manager
Business monitors customer retention outlook & performance
12 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
Demo FlowInsights from Data Science to Production
Data ScientistsWhere is the data I need to answer the business questions?
Business Users Where is the insight & predictions from the data?
HDP Cluster
Kno
x
AdminsHow do I meet SLA, Performance, .., Feature needs?
13 © Hortonworks Inc. 2011 – 2016. All Rights Reserved
What Have We Learnt
It Wasn’t My Fault
Model Data Lifecycle
Collaborative WorkSpace
Bring Your Own Tool
Built-In Security
Real Time Monitoring