Insurance

An insurance group consists of 10 property and casualty insurance, life insurance and insurance brokerage companies. The property and casualty companies in the group operate in a 17-state region. The group is a major regional property and casualty insurer, represented by more than 4,000 independent agents who live and work in local communities through a six-state region. Define the metrics to analyse agent performance based on several attributes like demography, products sold, new business, etc. The goal is to improve their existing knowledge used for agent segmentation in a supervised predictive framework.

Inspiration

Vizualization:

a. Summary stats by agency

b. Product line: Commercial Line/Personal Line wise analysis

c. Agency wise - state wise - distribution of Retention ratio (Top 10)

d. Quote system wise hit ratio distribution (also by PL/CL)

e. Add few more based on your understanding of the data

• Growth rates from 2006 to 2013 have to be computed and converted to independent attributes and include in the data for modeling.
• Compute Hit ratio based on bounds and quotes for each Quotes system
• Compute required aggregations at Agency id and state and year
• Decide if binning the data works for this situation

Some suggested approaches:

a. Model Building - Either Regression or classification

b. Pattern extraction - Classification Model

c. Patterns from the data using Decision Trees