Eliminating Bottlenecks in Analytics
I like buzz words – they arrive like a newly appointed passport to “coolness” and if you successfully incorporate them into your every day conversation, then you too can inherit that cool factor. The only problem is that they can quickly become over exposed to the point that every one gives you the “that is so last week” glare.
Yet some survive the test of time and frequency of use – maybe because they truly reflect a situation that is sometimes hard to describe. “Bottleneck” is in my opinion one such word. By the very definition, bottleneck implies a hazard that is hard to overcome. An area of congestion that can be measured in wasted time and inhibits a deliverable goal.
One such example of a known business process bottleneck is the gap between development and deployment environments and how to transpose the artifacts created using development technologies (that are typically off line) over to the deployment technologies (that typically execute real time).
We see this quite frequently in financial services organizations within their risk and analytics organizations. There is a significant gap between the analytics technology, used to develop new risk artifacts like attributes and models, and the production environment that executes them, such as origination platforms. While sophisticated development platforms enable the crunching of large quantities of data and the utilization of highly complex statistical algorithms, the end result often has no direct way of being consumed by the production platforms that need to execute the algorithms.
This represents the bottleneck where paper-based requirements and resources are used to bridge the gap between the two worlds of development and deployment and re-coding and re-testing is an acceptable solution to this breakdown in communication.
Equifax is helping customers by delivering solutions that helps bring together the analytical and production system and bypass IT bottlenecks. We enable attribute and models generated from a development system, such as SAS or R, to be easily consumed by the execution system with little to no re-coding and testing. Benefits are realized through the speedier implementation of attributes and models which in turn enables a faster time to market of key strategies. As model code can be easily transposed between the production and development platforms, recognizing the efficiency of a model and reacting to any modifications that are required can also occur in a much more stream-lined and efficient manner.
As both a provider and consumer of attributes and models, Equifax is not only committed to solving this bottleneck for internal efficiencies, but for our customers as well.
Recommended For You
Last week, top data and analytics experts from Equifax and around the world convened in Edinburgh, Scotland for Credit Scoring and […]
Adaptive AI is the next great advancement in leveraging AI for credit risk. Equifax’s Chief Innovation Architect, John Fenstermaker, developed the product […]
Equifax has developed Insight Score™ for Personal Loans, a risk score optimized to help lenders evaluate applicants seeking unsecured personal loans. I […]
In November, Equifax and Moody’s joined forces to recap the economic and credit trends of 2018 — and look ahead […]