Interactive Use Case Explorer for DREAM-KG

The team created an interactive use case explorer that documents 18 questions users might want to ask DREAM-KG. The explorer shows which queries are possible, which return partial results, and which return zero results. Users can filter by predicates, and the actual SPARQL queries are embedded with live links to run them on frink. The team identified missing entries and predicates that could improve DREAM-KG. They found that missing cities data can be solved by reverse geocoding. A Python script was created to perform reverse geocoding and output a CSV file with city, postal code, and full address for every service.

The team used AI to create the interactive use case explorer and an exploration report. They built a Python script that performs reverse geocoding to fill data gaps in the knowledge graph. All work was uploaded to a GitHub repository, including the Python script, CSV file, exploration report, and explorer code. The queries use SPARQL and link to frink for live execution.

The explorer helps the community understand data gaps and enhancements needed in DREAM-KG, a knowledge graph that connects homeless individuals with resources. The work provides concrete challenge ideas that other resource groups can address to improve the system.