Phnom Penh Urban Park Thermal Dataset: LST, NDVI, and Cooling Metrics from OpenStreetMap Polygons (2022–2024)

Description

Phnom Penh Urban Park Thermal Dataset: LST, NDVI, and Cooling Metrics from OpenStreetMap Polygons (2022–2024)Asian cities.Phnom Penh Urban Park Thermal Dataset: LST, NDVI, and Cooling Metrics from OpenStreetMap Polygons (2022–2024) This dataset provides a polygon-level thermal performance assessment of urban green and open spaces in Phnom Penh, Cambodia, derived from remote sensing and geospatial analysis. The dataset includes 128 OpenStreetMap (OSM) leisure polygons (≥0.3 ha) and integrates multi-source satellite observations with spatial metrics to quantify urban park cooling performance in a tropical secondary-city context. Land surface temperature (LST) is derived from Landsat 8 and 9 Collection 2 Level-2 products, composited across two dry seasons (November–March, 2022–2024). Vegetation density is measured using Sentinel-2 NDVI composites aligned to the same temporal window. For each polygon, the dataset reports mean LST, surrounding buffer LST (300 m), cooling intensity (CI), and cooling distance (CD), along with structural attributes including area, shape index, and compactness. Cooling intensity (CI) is defined as the temperature difference between the surrounding buffer and the polygon interior, while cooling distance (CD) captures the spatial extent of measurable cooling effects using a 30 m ring-based approach. The dataset also includes NDVI-based subgroup classification and derived indicators used in regression and spatial autocorrelation analyses. This dataset enables reproducible analysis of urban thermal dynamics in data-scarce environments and supports comparative studies of green infrastructure performance in rapidly urbanising Southeast Asian cities.

Steps to reproduce

0. Environment setup Create the Python environment using the provided configuration file: conda env create -f environment.yml conda activate park-thermal-env Alternatively, install dependencies manually: pip install -r requirements.txt Required Python version: 3.10+ 1. Data preparation (optional) The full analysis can be reproduced directly from the provided dataset (03_DATA/final/parks_analysis.csv). If rebuilding from raw sources is required, run: python 04_SCRIPTS/01_collect_lst_gee.py python 04_SCRIPTS/02_get_park_boundaries.py python 04_SCRIPTS/03_merge_lst_parks.py Google Earth Engine (GEE) GEE scripts are provided in: 04_SCRIPTS/gee/ A Google Earth Engine account is required Users must authenticate via: earthengine authenticate All asset IDs, temporal filters (2022–2024 dry seasons), and cloud masking procedures are explicitly defined in the scripts 2. Statistical analysis Run the full statistical pipeline: python 04_SCRIPTS/04_regression_analysis.py # OLS (HC3 robust SE) python 04_SCRIPTS/09_spatial_tobit_analysis.py # Tobit + Moran's I python 04_SCRIPTS/10_tobit_qr_analysis.py # Quantile regression 3. Recompute locked statistics Reproduce all reported numerical outputs: python 04_SCRIPTS/recompute_stats.py Output: 05_DOCUMENTATION/locked_stats_v7.json This file contains all summary statistics and regression results reported in the manuscript. 4. Manuscript reconstruction (optional) Rebuild the manuscript from source files: python 04_SCRIPTS/build_v7_docx.py Output: 01_MANUSCRIPT/KH-01_manuscript_v7_IJUS.docx 5. Reproducibility verification All numerical results in the manuscript are reproducible from the provided dataset and scripts Running Steps 2–3 must reproduce locked_stats_v7.json exactly Any deviation indicates environment or dependency inconsistencies 6. Data structure 03_DATA/ raw/ # satellite-derived inputs and OSM extracts processed/ # intermediate merged datasets final/ # analysis-ready dataset (parks_analysis.csv)

Institutions

• Handong Global University

  Gyeongsangbuk-do, Pohang

Categories

Licence