Synthetic European road freight transport flow data based on ETISplus

Published: 24 September 2021| Version 1 | DOI: 10.17632/py2zkrb65h.1


This dataset describes estimated European truck traffic flows between 1,675 regions all over Europe and is based on the publicly available ETISplus project from 2010 (DOI: 10.13140/RG.2.2.16768.25605). The project collected Europe-wide freight volumes and calibrated the resulting origin-destination matrices with real world traffic flows. For the current dataset, the truck results of the ETISplus project were updated using current Eurostat data ( Additionally, a forecast was added for 2030. Using Dijkstra's algorithm, the freight flows were finally transferred to the European highway network. Therefore, the dataset provides a synthetically generated truck traffic volume for each road section. The dataset can be a basis for developing, planning and sizing future road infrastructure, such as charging infrastructure for electric trucks. The dataset consists of four files: 01_Trucktrafficflow, 02_NUTS-3-Regions, 03_network-nodes, 04_network-edges. All of them are stored as comma separated values with commas as column separators and dots as decimal separators. The main dataset 01_Trucktrafficflow describes 1,514,573 directed transport flows in fifteen columns: (1) ID origin region, (2) name origin region, (3) ID destination region, (4) name destination region, (5) shortest path in the modeled E-road network, (6) distance from origin region to the E-road network, (7) distance within the E-road network, (8) distance from the E-road network to the destination region, (9) total distance, (10) road freight flow in tons for 2010, (11) road freight flow in tons for 2019, (12) road freight flow in tons for 2030, (13) truck traffic flow in number of vehicles for 2010, (14) truck traffic flow in number of vehicles for 2019, (15) truck traffic flow in number of vehicles for 2030. 02_NUTS-3-Regions contains a list with the regions under investigation. 03_network-nodes and 04_network-edges illustrate the highway network. The first contains the following information on each network node as columns: (1) node ID, (2) longitude of the location, (3) latitude of the location, (4) ID of the corresponding NUTS-3 region, (5) country code. The second contains information on the edges: (1) edge ID, (2) information whether the edge is manually added or part of the original ETISplus dataset, (3) length of the edge, (4) ID endpoint A, (5) ID endpoint B, (6) number of trucks in 2019 (both directions), (7) number of trucks in 2030 (both directions).


Steps to reproduce

The European Transport policy Information System (ETISplus) (DOI: 10.13140/RG.2.2.16768.25605, data: ) serves as a basis for the described dataset. In a first step, based on Eurostat data (, the national and international growth of freight transport volume was determined on a country-by-country basis between 2010 and 2019. Subsequently, the country-specific growth rates were used to scale the ETISplus data from 2010 to 2019. Since projections for 2030 vary extremely, the same growth was assumed through 2030 as in previous years. Afterwards, the freight volumes were then converted into vehicle trips using an average loading factor of 13.6 t and an empty trip share of 25%. The road network is also based on the ETISplus project. The original network was filtered for highways, four-lane roads and smaller roads, which are part of the European road network. To ensure that all E-roads are part of the dataset, a comparison with the current E-road network was done by hand and missing edges were manually added. The routes relevant for long-distance traffic, where public refueling or charging infrastructure will be increasingly needed in the future, are thus mapped. Then, the vehicle trips were mapped to the road network. Using Dijkstra's algorithm implemented in the Python library NetworkX, shortest paths between die origin and destination regions were determined. Additionally, regional traffic was excluded, since the regional resolution is not high enough to represent them properly. The distance within the region, which cannot be estimated cleanly, would be higher than the distance in the defined road network. Therefore, routes where origin and destination are in the same region were excluded. In addition, routes that do not have a network node in the origin or destination region and are less than 50 km apart or are directly adjacency were deleted. Finally, for each individual edge in the network, it was calculated how many trucks would use it.


Fraunhofer-Institut fur System und Innovationsforschung


Transport Infrastructure, Road Transportation, Infrastructure, Road Freight, Road Network