Advanced Business Analytics Seminar

Print Agenda

Business analytics has become a major force of innovation in the insurance industry and beyond in the past decade. Inexpensive computing power, rich data sources and increasingly powerful analytical tools become more widely available every year. As a result, quantitatively skilled professionals have unprecedented opportunities to bring scientific rigor to areas of strategic importance within their organizations. For actuaries, the business analytics revolution affords opportunities to ground traditional actuarial work in rigorous statistical methodology and apply their skills in new domains.

This interactive, hands-on seminar will impart practical working knowledge of statistical and machine learning techniques that are relevant in actuarial work. Core techniques such as regression analysis, generalized linear models, generalized additive models and regularized regression techniques will be covered. Additionally, survival models, decision tree analysis and such “unsupervised learning” techniques such as principal components analysis and clustering will also be discussed. This course will include a blend of brief lectures complemented by stylized hands-on data analysis case studies amenable to classroom treatment.

The seminar will combine theory and practice with theoretical discussions that emphasize concepts and intuitions rather than mathematical formalism. These theoretical discussions will enable attendees to better interpret, communicate and critically examine model outputs. Practical working knowledge of data analysis methodology will be imparted through a series of data analysis case studies. Seminar participants will receive datasets and computer code to work through examples with the instructor.

Inherent to modern statistical and actuarial practice is the ability to compute with data. To that end, the medium of the seminar will be the open-source R statistical computing environment. R is widely considered a primary working language of modern data analysis, and provides users with exceptional facilities for manipulating, graphically exploring and modeling complex datasets.

Attendees will receive a self-contained introduction to R package prior to the seminar that explains how to install and begin using R. Working knowledge of R will be further cultivated during the workshop as participants analyze data in real time.

At the conclusion of this seminar, participants will be better able to:

• Perform basic data manipulations and fit a variety of standard models in the R statistical computing environment
• Graphically explore data to motivate various modeling choices and graphically criticize models and motivate model improvements
• Interpret and critically examine standard model output
• Test the performance of models on holdout data
• Translate a business problem into the design of a data analysis strategy

This seminar will convey foundational information on this topic and a general understanding of statistics is the only prerequisite. The seminar will focus on gaining theoretical intuition as well as practical experience with the “art” of interactively exploring data and building, validating, and interpreting predictive algorithms.  It is designed to be accessible to actuaries with little or no experience with predictive analytics. Given that a working knowledge of applied data analysis in R cannot be gained in only a few days, the goal of the seminar is to convey a theoretical framework and computational toolset sufficient for participants to continue to gain expertise on the job. To create an optimum learning environment, this seminar will be limited to 35 registrants.

Seminar Instructor
James Guszcza, FSA, FCAS, Ph.D.
US Chief Data Scientist
Deloitte Consulting

James Guszcza, FSA, FCAS, Ph.D., is Deloitte Consulting’s US Chief Data Scientist and a member of Deloitte’s Data Science practice. He is a fellow of the Society of Actuaries (SOA) and the Casualty Actuarial Society (CAS) and has a Ph.D. in the philosophy of science from the University of Chicago. Guszcza spent the 2011–12 academic year as an assistant professor in the actuarial science, risk management and insurance department of the University of Wisconsin-Madison. A frequent author and conference speaker, he also designed the CAS Predictive Modeling Limited Attendance Seminar.