Tutorial: Using R for Customer Analytics

presented by Jim Porzak

This tutorial will be very down-to-earth. Actual customer data (sanitized, of course) will be used and provided to the attendees so they can work through the examples demonstrated. All techniques and methods discussed are actually used by Loyalty Matrix to provide actionable insights to our clients.

While R has achieved wide acceptance in the academia, it has yet to reach that status in the business community. There are a number or reasons for this. One is the lack of a business orientated introduction. This tutorial provides a gentle introduction to R and surveys the many ways R can extend the modern business analyst's tool set.

Only the most basic R skills are assumed.

Detailed outline:

1. Introduction:
   

   • What is "customer analytics" and why do we do it?
   • Implications of working in a business environment.
   • Specific Loyalty Matrix tools & biases.

2. Part I - Getting Started:
   Brief review of what needs to be done before serious analysis can start.

   • Sourcing business requirements.
   • Sourcing raw data.
   • Profiling raw data.
   • Data quality control & remediation.
   • Staging data for analysis.

3. Part II - EDA and Basic Statistics:
   Step-by-step look at basic customer data with three important variations of the usual business model.

   • The fundamentals: counts, amounts and intervals.
   • The geographical view.
   • Subscription businesses.
   • Hospitality businesses.
   • Big ticket businesses.

4. Part III - Mining, Modeling, Segmentation & Prediction:
   In-depth look at some useful packages for advanced customer analytics.

   • Decision tree methods - rpart, tree, party and randomForest.
   • Survival methods - survival and friends
   • Clustering methods - mclust, flexclust.
   • Association methods - arules.

5. Conclusion:

   • Review of applicable methods by type of client.
   • The customer analytics check list.

Back to Tutorials