Use cases
One store for your knowledge graph, and for the networks on the ground
TripleTrail serves any RDF knowledge graph. Where your data describes rails, roads, pipes or cables, it also answers where each asset is and how it connects, questions that usually need two or three other systems.
Knowledge graph teams
A fast, modern triple store, map or no map
Most knowledge graphs have no geometry at all, and TripleTrail is built for them too. It leads the BSBM explore and update workloads among the engines we tested, replaces Jena Fuseki without client changes, and serves the RDF4J REST API, so RDF4J and GraphDB clients connect with only the URL changed.
Standards first
SPARQL 1.1 and 1.2, Graph Store Protocol, SHACL, SHACL-AF, RDFS and OWL 2 RL, RDF canonicalisation. Your queries and shapes stay portable.
Data in
SPARQL Anything-compatible Facade-X queries and R2RML/RML mappings over files and databases, built into the server.
Data out
Linked Data pages with content negotiation, a JSON-LD REST API generated from your shapes, full-text search and GeoParquet export.
Railways
An infrastructure register you can ask anything
Infrastructure managers publish and exchange data as linked data, for example the EU’s Register of Infrastructure (RINF). TripleTrail loads it as it is and adds what a register alone cannot do.
- Find every signal, switch or platform near a point, along a line or in a region
- Turn kilometre posts into map positions and back (linear referencing)
- Route over the real track topology, with rules for gauge, electrification or train length
- Check the data with SHACL shapes before it is published
- Lift spreadsheets and exports from asset systems into the graph, on a schedule
SELECT ?seq ?edge ?agg_cost WHERE { SERVICE <http://w3id.org/tripletrail/pgr> { [] pgr:function rail:route ; pgr:network ex:mainLines ; pgr:start_vid ex:stationA ; pgr:end_vid ex:stationB ; pgr:constraint "?c_gauge = 1435 && ?c_electrified && ?c_axleLoad >= 22.5" ; pgr:seq ?seq ; pgr:edge ?edge ; pgr:agg_cost ?agg_cost . } }
Roads
Road networks with every restriction in the data
Weight limits, height limits, one-way streets and closures are facts in the graph. TripleTrail routes with them directly, instead of exporting a routing graph every night.
- Routes per vehicle type, with SHACL rules for which vehicle may use which road
- Travel-time areas: what can a crew reach within 30 minutes?
- Detours around works, with the assets along the way
- Serve the network to QGIS and web maps with OGC API Features and vector tiles
5 ms
for a 50 km driving-distance area on a 750,000-edge network.
0.65 s
for a 100 × 100 travel cost matrix, against 25 s in pgRouting.
Any model
Route over your own vocabulary. No fixed table layout.
OGC
QGIS connects to TripleTrail like to any feature server.
Utilities and energy
Your grid model already is a graph
Electricity grid models in IEC CIM and CGMES are exchanged as RDF. TripleTrail adds location and connectivity, so you can ask questions that cross the model and the map.
- Which customers sit behind this substation, and where are they?
- Which parts of the network fall apart if this cable is cut (connected components)
- Which assets lie in a flood zone, a works area or a protected area
- Water and telecom networks work the same way
From files to a living model
Lift CSV, Excel, Shapefile, JSON and XML into the graph with SPARQL Anything-style queries or RML mappings, validate it with SHACL, infer what follows with rules, and publish it as an API. One pipeline.yaml runs the whole chain, in the Workbench or headless in CI.
- lift → load → update → infer → validate → export
- Regression tests compare every step with the last release
- Canonical digests show exactly what changed between two releases
Is your network not on this page?
If your data has a place and a connection, TripleTrail probably fits. Tell us what you manage.