High-Resolution Mapping of Vacant Housing in China Using Night Lights and POI Data

A dataset containing source code and sample data for estimating building-scale housing vacancy rates across mainland China. The method fuses NPP-VIIRS night-time light intensity, OpenStreetMap POI density, and road network density into a composite human activity index to identify potentially vacant residential buildings without relying on household surveys. The dataset includes preprocessing scripts, a complete estimation workflow, and a case study dataset to demonstrate the process.

Use Cases

• Estimate housing vacancy rates at the building scale based on fused multi-source spatial data.
• Calibrate local activity thresholds for vacancy detection at the county level.
• Batch process housing vacancy estimations across provinces using the provided workflow.
• Identify potential ghost towns or underutilized urban areas based on low human activity indices.

Strengths

• Includes a complete, documented workflow for housing vacancy rate estimation.
• The method leverages multiple open data sources (NPP-VIIRS, OpenStreetMap) for robust inference.
• Provides a case study dataset to demonstrate the full analysis pipeline.

Limitations

• Row count and specific column definitions are unknown, limiting suitability assessment.
• The data's freshness and temporal coverage are not specified in the provided metadata.
• Actual data quality and coverage require manual inspection after downloading the 163.5 MB ZIP file.

Provenance

Source

figshare

Collection Method

Multi-source spatial data fusion of satellite night lights, POI, and road network data.

Freshness

Last updated 2026-05-12 09:16:34.

Geography

Mainland China