Skip to content

mie-lab/velonemo_v2

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

VeloNEMO v2 — Scaling Bikeability Knowledge Extraction with LLMs and Geospatial Knowledge Graphs

Status: early development — no stable code or releases yet. Structure and APIs will change.

VeloNEMO v2 is the scaled successor of VeloNEMO, a formal ontology for harmonizing bike network evaluation metrics (Grisiute et al., 2024). Where v1 was a proof of concept built from a manually curated corpus of 25 papers, v2 extends the knowledge model and populates it automatically from a large heterogeneous document corpus using a schema-constrained LLM extraction pipeline.

What v2 extends

The knowledge model grows along three axes:

  • Evaluation perspective: explicit modeling of objective vs. perceived evaluations, replacing v1's Perceived name-prefix.
  • Decision structure: an index's modelling commitments (metric orientation, normalization, weighting, compensation, spatial interdependence) as queryable entities.
  • Provenance: geographic and temporal context of each evaluation, so commitments can be analyzed as spatio-temporally situated choices.

The model is designed in two layers: a generic mid-level model for composite urban indices, and a cycling-domain specialization. The same structure is intended to extend to walkability, liveability, and related index families.

Planned repository structure

ontology/     VeloNEMO v2 (OWL) + validation shapes
extraction/   schema-constrained LLM extraction pipeline
resolution/   entity resolution and canonicalization
validation/   agreement scoring against the v1 ground truth
analysis/     queries and corpus-scale analyses

Setup

conda env create -f environment.yml
conda activate velonemo2

Related work

Contact

Ayda Grisiute — Institute of Cartography and Geoinformation, ETH Zurich

About

scaling velonemo ontology

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages