Data

to view this content, you must download the flash player.

You can't learn something from nothing.

While carriers often understand the importance of attention to fundamentals when building a predictive model, they often fail to recognize the importance of data integrity and richness at the start. Valen's Data solution focuses on optimizing a customer's data set and then enriching it with data from Valen's External Data Warehouse (EDW). The end result is a high-quality, statistical-grade data set, optimized for use in modeling and analytics.

Data Hygiene, Cleansing & Validation

Valen's solution checks data integrity to ensure that the data set does not contain keying errors, referential integrity errors and/or nonstationary issues. If any of these occur, a suboptimal model will result. Manually analyzing thousands of pieces of data takes months of dedication and time - something most market-driven companies do not have. Furthermore, by the time this entire process is manually completed, much of the data already may be out-of-date - largely defeating the purpose of the exercise.

Automated data validation drives the focus of Valen's automated hygiene and cleansing efforts. Using a highly automated approach, the hygiene work produces final client data sets, which typically contain over 95% of the original data analyzed. This process not only creates a high-quality foundation for Modeling & Analytics, but saves carriers valuable time in the overall modeling process.

Data Enrichment

When building predictive models, carriers typically rely on data from their own policy administration and claims systems. This means that most carriers are leveraging relatively the same standard information, and are likely to derive the same insights as the competition. But preparing your internal data is just the first step. Carriers that combine internal data with meaningful external data can gain unique insights that competitors may not have the ability to see. By accessing additional data from external sources and then linking them to internal data, and considering variable interactions from among the different sources, more impactful models are developed. Ultimately, this can provide carriers with many means to outmaneuver the competition.

External Data Warehouse

The Valen External Data Warehouse (EDW) is a constantly growing repository of attributes consisting of hundreds of publicly available data elements as well as some purchased elements. By supplementing client data with additional data variables from Valen's EDW, Valen can develop predictive models that segment risks at a much finer granularity. This micro-segmentation results in more accurate decisions for our customers, often in the form of outselecting the competition.

While traditional external data sources have been utilized for some time in risk analysis, powerful new data sources are emerging and can provide significant new insight when combined with internal data. Information such as distance to the nearest hospital or police station, or safety information about transportation devices, can help carriers segment their book with finer granularity.