Kosovo Road Surface Data: 169,100 km of AI-Augmented Infrastructure

HeiGIT provides road surface classifications for 169,100 km of roads in Kosovo, combining OpenStreetMap (OSM) data with AI predictions from Mapillary imagery. The dataset distinguishes between paved and unpaved surfaces, addressing a gap where 52.9% of OSM surface tags were previously missing. It was last updated in March 2026.

Use Cases

• Infrastructure planning using pred_class to identify unpaved segments for maintenance
• Climate emissions modeling based on surface and smoothness attributes
• Logistics analysis using highway and urban flags to assess rural accessibility

Strengths

• Covers 169,100 km of road network
• AI-derived predictions for segments with missing OSM tags
• Includes urban classification from GHSU and AFRICAPOLIS layers

Limitations

• High percentage of missing surface data (52.9%) in the base OSM layer
• AI predictions are limited by the availability of Mapillary street-level imagery

Provenance

Source

HeiGIT (Heidelberg Institute for Geoinformation Technology)

Collection Method

Hybrid deep learning using Mapillary street-level imagery combined with OpenStreetMap (OSM) and urban layers.

Freshness

Last updated March 2026.

Geography

Kosovo