Are you interested in the intersection of Business Intelligence and the Semantic Web?
Benchmarking an OLAP engine on Linked Data
- Decision support over Linked Data requires typical Business Intelligence queries such as "The top 10 most discussed product categories of products from a specific country based on number of reviews by reviewers from a certain country" or "Find 10 cheapest vendors for a specific product type by the ratio of products below and above the average."
- Triple stores have been built that can ask such queries using SPARQL
- However, decision makers do not issue such SPARQL queries. Rather decision makers use interfaces to analyse the data which in turn issue these SPARQL queries. One common interface in business is provided by OLAP client tools.
- OLAP client tools work on OLAP engines and issue typical OLAP queries or operations such as slice/dice or drill-down/roll-up.
- Olap4ld  is an open-source OLAP engine on Linked Data that allows decision makers to issue such typical OLAP queries
- In this work, olap4ld will be benchmarked against a certain dataset and queries
- The dataset and queries that will be used are provided by the Berlin SPARQL Benchmark (BSBM) - Business Intelligence Use Case 
- The challenges of this thesis are
- Effectively transforming the BSBM dataset so that it can be queried by olap4ld. Olap4ld is using the standard RDF Data Cube Vocabulary .
- Reformulating the BSBM queries into common OLAP queries that are understood by olap4ld. Olap4ld is implementing the Open Java API olap4j that is using the standard OLAP query language MDX.
- Adaption of the BSBM (Berlin SPARQL Benchmark) suite for benchmarking to run a benchmark for olap4ld
- Within this thesis the student will gain a good understanding of OLAP on Linked Data. At the end of the thesis, the student may suggest improvements towards faster queries in olap4ld.
- Experiences with Java (not on the UI side)
- Some knowledge about Semantic Web or Linked Data