Pseudo-space Learning Data: Urban Heat Exposure Risk for Australian Statistical Areas

PSL model provides a 2.1 GB repository of data and code for predicting urban heat exposure risk using pseudo-space learning. The workflow integrates Google Earth Engine, QGIS, and R to process remote-sensing data, building footprints, and census population data for Australian statistical areas (SA1 and SA2). The repository was last updated on 2026-05-03 and is licensed under CC-BY-4.0.

Use Cases

• Train machine learning models for urban heat island prediction based on integrated remote-sensing and census data.
• Apply regression kriging for spatial interpolation of environmental risk factors.
• Reproduce scientific workflow for generating heat exposure maps using the provided R and Python scripts.
• Conduct downstream spatial visualization and analysis in QGIS using the provided shapefiles.

Strengths

• Repository is 2.1 GB and includes code, raw data, model outputs, and visualization files for full reproducibility.
• Workflow integrates multiple tools (Google Earth Engine, QGIS, R) for a comprehensive spatial analysis pipeline.
• Data sources include raw remote-sensing data, building footprints, and census population data.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Geographic coverage is limited to Australian statistical areas.

Provenance

Source

PSL model via figshare.

Collection Method

Processed from raw remote-sensing data, building footprints, and census population data using Google Earth Engine, QGIS, and R.

Freshness

Last updated 2026-05-03 02:55:00; freshness should be verified.

Geography

Australia (covers SA1 and SA2 statistical areas).