A Benchmark Dataset of 13 Routes with Semantic Annotations Mapped onto a Mobility Network

These datasets constitute a collection of reference datasets used to evaluate a data model for the semantic annotation of routes. The benchmark is composed of a set of routes, the segments of the mobility network through which these routes pass, and a set of aligned semantic annotations associated with the routes.

The 13 routes, derived from diverse collaborative sources, are stored in the file routes.csv. The geometries of the routes were map-matched to the BDTOPO road network representing the mobility network and are stored in the mobilite layer of the GeoPackage. The map-matching process is described in: https://doi.org/10.1080/23729333.2019.1615730.

The semantic annotations of the routes are also stored in the GeoPackage in the following layers:

pii_balcons for route segments running along balconies,

pii_lacet for route segments consisting of a series of hairpin bends,

pii_foret for segments running along or through forests,

pii_lac for route segments running along lakes,

pii_pont for segments crossing bridges.

This archive is released for transparency and reproducibility purposes.