Multi-Relational Urban POI Graph Dataset for Baku Derived from Overture Maps

Read full text
Authors: Etibar Vazirov, Simone Monaco, Daniele Apiletti
Cited by: 0

Abstract

This dataset provides a multi-relational urban Points-of-Interest (POI) graphconstructed for the city of Baku (Azerbaijan) using data derived from Overture Maps.The dataset includes processed parquet files containing spatial coordinates,categorical attributes, and region-level information required to build aheterogeneous urban graph suitable for graph-based anomaly detection andrepresentation learning. The graph construction follows a spatial-semantic relation design includinggeographic proximity, shared categorical properties, and regional grouping.These relations enable the study of embedding-based unsupervised anomalydetection, stability analysis across random seeds, and spatial characterizationof consensus anomalies. The dataset accompanies the research work: “Embedding Stability and Consensus-Driven Anomaly Discoveryin Multi-Relational Point-of-Interest Graphs”. All files are provided to ensure transparency, reproducibility, and reuse infuture urban graph learning and anomaly detection studies.

Fields of Study & Subjects

Advanced Graph Neural Networks Anomaly Detection Techniques and Applications Geographic Information Systems Studies

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
16 February 2026
Language
EN
Document type
Dataset