A Unified Framework to Combine Sundry Data Sourcesby Michael Murff
Head, Risk Science Programs
Data scientists require robust computational systems that scale as data grows. Frequently, data systems become fractured as new stores come online, and as technology changes. Currently, data users find data in the EDW, assorted data marts, and flat files. This means an inordinate amount of time is spent on data preparation toward dataset creation for modeling, analysis, and production. Thus, there is a need for a unified framework to combine sundry data sources, enable analytical processing, and channel rapid deployment to production. We devise a Hadoop/MR product solution aptly named: System for Profiling of Entities & Analysis of Relations.