Scoring Methodology

 

The Score Data diagnostic calculates a Data Quality Score and an Overall Risk Score for individual locations and for entire portfolios. The scores help you identify high-value exposures that warrant further investigation and augmentation (enhancement).

The scores are peril-specific. The impact that missing exposure data has on modeled losses depends on the location and characteristics of the risk and the peril under consideration. For example, missing construction characteristics have a greater impact on modeled losses in areas where there is a lot of heterogeneity in the types of construction used and when the scoring methodology is considering a peril for which construction is a key variable.

Data Quality Score

This score measures the completeness of the exposure data; it quantifies inaccurate and missing primary risk characteristics in the exposure data that can contribute to uncertainty in modeled losses. Ranging from 0-100, a higher data quality score indicates better relative data quality, which can lead to reduced uncertainty. The application calculates data quality scores for each location in the analysis target, and then combines them to produce various aggregate scores, such as for the entire portfolio and for each state. Portfolio-level data quality scores are particularly valuable because they can help you to

         Communicate portfolio-level data quality to underwriters, brokers, reinsurers, and rating agencies

         Compare exposure data quality in portfolios

         Assess the impact of a Data Quality Augmentation

         Monitor data quality enhancement of a portfolio over time

When calculating data quality scores for individual locations, Data Quality considers the availability of known data for the following primary risk characteristics:

         Geocode match level (Relaxed Address or better indicates good quality address information)

         Building construction code

         Occupancy code

         Year built

         Number of stories

Overall Risk Score

This score combines Data Quality scores, region- and hazard-specific risks, and replacement values to identify exposures with high values, high risks, and poor data quality. At the location level, overall risk scores are represented as monetary values; they can be very large because they are a direct function of the replacement value. Higher scores indicate high-risk locations that may be key areas of concern. At the portfolio level, overall risks are calculated as a weighted average of all the locations in the portfolio.

The overall risk distribution is divided into seven bands, or "categories," where each range represents the number of locations whose overall risk falls within the range of that category. The minimum is "0," and the maximum of the range depends on the overall risk of the largest and most vulnerable location in your exposure view.

Overall risk scores can help you to

         Prioritize portfolios with high value, high risk, and poor data quality

         Prioritize individual locations or subsets of locations for validation or augmentation

For Policy and Geography views, the Data Quality Score results include a Data Quality Uncertainty value that reflects the amount added to the overall risk score because of deficiencies in the exposure data.

 

© 2016 AIR Worldwide. All rights reserved.

Click here for information about AIR Worldwide, including copyright, trademarks, confidentiality, and contact information.

Touchstone V3.0 Updated December 01, 2016