As an insurance company, we receive a monthly premium from policy holders and in return, we pay claims on veterinary bills. Insurance risk for pet health is relatively uncharted territory; identifying key patterns can affect the company in a big way.
As members of the Analytics department, the audience for our work is often the rest of the company. Our projects are not always requested and our solutions might call for a change in workflow. With predictive modeling being a relatively new concept to us, effect communication is as important as validity of results.
In this talk, we will briefly walk through setting up problems in analyzing customer churn and claims projection. Datasets are generated and structured using R; models are built using DataRobot; results are communicated using Excel.
After two years working in financial planning & analysis and Milliman’s healthcare practice, TJ joined Trupanion while studying accounting, finance and math. Since joining full time, he’s worked in several roles at Trupanion in operations and regulatory & pricing. He’s now leading business intelligence, which satisfies his intellectual curiosity as he mines Trupanion’s data for information to assist the company’s operations and provide value to our customers. TJ has a Bachelor’s in Business Administration in Accounting from Gonzaga University and is near an ASA.
BS in mechanical engineering from the University of New Mexico
– Studied machine learning with a focus on applications in robotics
– Started work as a Data Analyst at Trupanion March 2014
– Responsibilities include call center routing optimization and claim outcome prediction