Seminar on Computational Learning and Adaptation


  Real-world Insights from Mining Retail E-Commerce Data

Ronny Kohavi
Blue Martini Software
San Mateo, California
http://www.bluemartini.com/bi
http://robotics.stanford.edu/~ronnyk/

The Business Intelligence team at Blue Martini has been doing data mining of e-commerce data for several clients over the last three years. We will briefly discuss the system architecture that lets us collect, transform, and mine data effectively, then show examples of insights from real customer data. The examples are related to usability, bot detection, session timeouts, search, referrers, micro-conversion rates, product affinities (associations), campaigns, RFM analysis, and geographical analysis. We will even show a real example of Simpson's paradox that we encountered. The talk will be mostly non-technical and hopefully fun.

In next week's seminar, Rajesh Parekh from Blue Martini will present lessons and challenges based on our experience.



Date: Thursday, May 22

Time: 4:15-5:30PM

Place: Cordura 100


Return to the seminar schedule