Semantic Data Integration
In the current state of technology, applications query sets of pre-defined data sources. When a data schema changes, application logic (e.g. business rules) and presentation logic must change. This brittle architecture delays the delivery of business capabilities, and greatly increases labor costs. Further, queries are written to specific databases; therefore the entire set of available databases and other content repositories are not considered. Although the answer to a basic question may reside in the “system”, the query will never know. There exists vast, untapped knowledge potential in stand-alone databases and data sources. When viewed separately, these databases each possess intrinsic value to those who understand them.
But when linked together with other databases, the value increases dramatically.
The lack of common data models presents a significant challenge in approaching this problem. We overcome this challenge by using an agile framework for describing data--a framework that easily integrates data from many sources, and accommodates unforeseen changes.
For this project we combined data from multiple sources and formats (various local, state and law enforcement databases) to realize a unified information model. The unified information provides law enforcement analysts (and machine analytics) the ability to interpret multi-source, multi-format and multi-domain data. Analysts can then query and modify the result (including rejecting and adding data) to create entity profiles. This capability also provides data lineage; every element of data may be traced to its origin without the need for a central repository.
We partnered with the Digital Enterprise Research Institute (DERI - www.deri.ie) to create and present this data integration capability. It allows adopters to integrate rapidly an arbitrary number of data sources, including cloud sources, using the W3C Resource Description Framework (RDF) standard.