This paper addresses performance challenges in the European Union Agency for Railways (ERA) Route Compatibility Check (RCC) tool, a system that intensively uses SPARQL to query a large-scale knowledge graph. The application determines if a vehicle type is compatible with a railway route’s characteristics, but a naïve approach of checking all route tracks leads to poor performance. To resolve this, we implemented and evaluated two performance improvement strategies: track aggregation and query parallelization. Track aggregation groups tracks with identical technical parameters to drastically reduce the number of evaluations during query execution, while parallel execution maximizes the throughput of the SPARQL endpoint. Our evaluation, comparing the optimised strategy against the naïve baseline, demonstrates a significant, approximately five-fold improvement in execution time for long routes. This work shows that combining data summarization with parallel execution is an effective approach for making SPARQL-intensive applications on large knowledge graphs performant and responsive.