Harmonized Annual Averaged Traffic Data at Street Segment Level for European Cities

Published in Nature Scientific Data: https://doi.org/10.1038/s41597-025-05698-y

Zenodo Archive:

This repository provides a framework for standardizing road traffic data from various European cities. Many cities offer road count data on their open data websites, but these datasets often differ in language, format, variables, and geometries. This project aims to harmonize these datasets into a standardized structure for ease of analysis and integration.

Repository Structure

Each folder in this repository corresponds to a specific city and includes the following components:

• raw/
  Contains the raw data downloaded directly from the city's open data website. If the original dataset is too large, a sample is provided to serve as an example.

• source.txt
  A text file containing the URL of the page where the raw data can be downloaded.

• {cityname}.ipynb
  A Python Jupyter Notebook that processes the raw data and converts it into standardized traffic data.

• treated/
  Contains the processed (treated) data in a standardized format.

Data Naming and Content

Processed data in the treated/ folder typically consists of GeoJSON files, named using the following format:

{cityname}_{variable}_{year}.GeoJSON

A summary of available processed data can be found in assets/cities_all.csv. Analysis of this table can be made in cities_available.ipynb.

Required Variables

• Average Annual Daily Traffic (AADT)
  Represents the average daily traffic volume for all days of the year.

• Average Annual Weekday Traffic (AAWT)
  Represents the average daily traffic volume for weekdays / working days.

Some datasets may include both variables, depending on what is provided by the city. The geometries in the final files are expected to be one of the following:

• Points: Representing the location of a road section.
• Oriented LineStrings: Representing road sections with directional flow.

Data Dictionary

Local Dataset

Feature Name	Type	Description
AADT or AAWT	float	Average Annual Daily (Weekday) Traffic (vehicles per day).
s85	float	85th percentile of observed speed (km/h).
savg	float	Average monitored speed (km/h).

Abbreviations

Abbreviation	Example	Meaning
TR	TR_AADT	Trucks (variables related to trucks/heavy vehicles).
2W	2W_savg	2-Wheels (variables related to 2-wheels).
pct	TR_pct_AAWT	Percentage (value as a percentage of AADT/AAWT).

Road Segment Information

Feature Name	Type	Description
geometry	Point / LineString	Geographic coordinates of the sensor location (Point) or the corresponding road segment (LineString) in the WGS84 projection.
raw_name	text	Name of the road segment.
raw_oneway	boolean	Whether or not the sensor monitors flow in one direction only. If this is not provided, one may look at the city-level labeling in cities_summary.csv.
raw_direction	text	Indication for the direction of the flow. This field can be an angle value or text indication (not translated). NaN if raw_oneway is False.
raw_lanes	float	The number of lanes of the monitored road segment.
raw_techno	text	The counting technique employed when this information is provided at the sensor level (not translated). Otherwise, one may look at the city-level attribute Table 1.

OpenStreetMap

Feature Name	Type	Description
osmid	float	The "road segment ID" that links the provided data to the OSM referential.
osm_name	text	Name of the matched road segment.
osm_type	text	The OSM “highway” classification type.
osm_lanes	float	The number of lanes of the matched road segment. Many values could be missing.
osm_oneway	boolean	Whether or not the matched road segment has flow in one direction only. Retrieved from the presence of the oneway attribute or lanes:backward/forward in OSM. Many values could be missing.
osm_maxspeed	float	The indicated speed limit on the road segment (km/h) retrieved from “maxspeed” in OSM.
osm_distance	float	The distance (meters) between the sensor location and the matched road segment (for point geometries only). This can serve as a quality indicator for the matching process.

How to Contribute

We welcome contributions to this repository! You can contribute by adding new cities and their corresponding data sources.

Installation

To set up the required dependencies, ensure you are using Python 3.12.8, then run: pip install -r requirements.txt

Contribution Guidelines

1. Include a sample of the raw data in the raw/ folder (if the full dataset is too large).
2. Provide the source URL in the source.txt file.
3. Include a Jupyter Notebook ({cityname}.ipynb) that transforms the raw data into the standardized format.
4. If the raw data files are too large to upload, include only the URLs, the processing code, and the final processed data.

Notebook Guidelines

To optimize the repository's size and usability:

• Avoid excessive figures or interactive maps in the notebook files to conserve space in the repository.
• Provide clear explanations within the notebooks, including:
  • Translations of variable names or key data terms.
  • Additional documentation when needed (e.g., external references or detailed metadata).
  • Justifications for any assumptions or transformations applied to the data. The files map_matching_OSM.py and temporal_agg.py can be used.
• The final processed files should be uploaded to the treated/ folder.
• Ensure the final data uses one of the accepted geometries (Points or oriented LineStrings).

This ensures transparency and reproducibility for all contributors and users.

Reference

Bonnemaizon, X., Ciais, P., Zhou, C. et al. Harmonized Annual Averaged Traffic Data at Street Segment Level for European Cities. Sci Data 12, 1365 (2025). https://doi.org/10.1038/s41597-025-05698-y