Data and Code for "Spatial patterns of traffic saturation on Flemish highways: a six-year time series cluster analysis"

This repository contains the datasets and data sources, analysis code, and workflow associated with the manuscript "Spatial patterns of traffic saturation on Flemish highways: a six-year time series cluster analysis". The following resources are provided: Data Files: road_segments_with_clusters.geojson: Geographic representation of 1044 highway segments, of which 819 with complete time series data and cluster assignments for workdays (6 clusters), Saturdays (3 clusters), and Sundays (4 clusters). Includes attributes: nat_ref (highway code), segment (segment name), and cluster label per day type. Coordinate system: EPSG:31370. traffic_saturation_timeseries.csv: Monthly traffic saturation time series for 819 highway segments covering January 2018 to December 2023 (72 months), with separate columns for workdays, Saturdays, and Sundays. Code: analysis.py: Python script implementing the complete analysis workflow, including: Data preprocessing (min-max scaling and 3-month rolling average) Dynamic Time Warping (DTW) distance computation Hierarchical Agglomerative Clustering with Ward's linkage Silhouette score validation Visualization of dendrograms and mean time series per cluster The script generates the main figures presented in the manuscript (silhouette plots, dendrograms, and mean time series plots with standard deviations). The analysis applies time series clustering to identify distinct spatiotemporal traffic saturation patterns across the Flemish-Brussels highway network. Traffic saturation is calculated as the intensity-to-capacity (I/C) ratio using Passenger Car Equivalents (PCE) from double inductive loop detectors. Analysis was conducted separately for workdays (Monday-Friday, excluding public holidays), Saturdays, and Sundays. Hierarchical clustering with DTW distance revealed: - 6 distinct clusters for workdays - 3 clusters for Saturdays- 4 clusters for Sundays Data Sources Agentschap Wegen en Verkeer (Flemish Traffic Control Centre) (http://indicatoren.verkeerscentrum.be/vc.indicators.web.gui/frontpage/index?page=%2Ffrontpage%2Findex.html). Data accessed: 2024. The dataset provided here includes preprocessed and clustered results; the original raw data remains available from the Flemish Traffic Control Centre. Reuse of the original data is permitted under the Open Data License of the Flemish Government (Gratis Open Data Licentie van de Vlaamse Overheid). Temporal coverage: January 2018 - December 2023. Geographic coverage: Flanders and Brussels Capital Region, Belgium. Highway segment geometry derived from OpenStreetMap (OSM), providing the spatial delineation of controlled-access highway segments used in the analysis. OSM data is made available under the Open Data Commons Open Database License (ODbL), which permits reuse with attribution.