Semantic Data Management for ad-hoc business reporting
Retrieving particular information from a vast number of business data sources is a challenge for data infrastructures of modern enterprises, as it commonly relies on functions and transformations that are statically provided by application developers. However, end users often request business reports in an ad-hoc fashion which involves operations over heterogeneous data sets which were not necessarily interconnected in the past.
One possible approach to improve this situation is to use Linked Data to connect multiple data sources on demand and to query it for particular information. In order to automatically answer a business-related query from an end user, a question from an input interface has first to be mapped to a corresponding query through the identification of relevant entities by using appropriate ontologies. This query can then be run on one or multiple data sources which provide an interface to access data and meta data in a Linked Data fashion, and return a result that can later be visualized to the end user in a frontend user interface.
The current project applies such an approach on exemplary business data of a financial department. Therefore, requirements have to be defined on the type of requested business data and on the business operations that have to be supported. A literature review has to compare already existing approaches in the context of information retrieval and semantic data queries. After that, an appropriate concept has to be formulated that describes both the entity mapping, query construction and query execution. A proof-of-concept implementation has to discuss the feasibility of the approach. Finally, the solution has to be evaluated based on the posed requirements based on a particular use case.