<b>A 1-km resolution dataset of affordability-constrained accessibility to elderly care facilities in China (v2)</b>

1. OverviewThis dataset provides 1-km resolution measurements of accessibility to elderly care facilities across mainland China under both geographic and affordability constraints. It is designed to support research on spatial planning, service equity, demand estimation, and policy evaluation in the context of population aging. Accessibility was calculated using an enhanced two-step floating catchment area method with economic affordability constraints. Therefore, it should be interpreted as a composite measure shaped jointly by service supply, population demand, and travel-time impedance, rather than as an indicator determined by any single component alone. 2. Main contentsThe dataset comprises five main components with clearly defined file names: Raster data (`1_Raster_Data/`): Files follow the naming convention `{BedType}_t{Threshold}_k{K}.tif` and `Gap_{BedType}_t{Threshold}_k{K}.tif` (e.g., `Total_t60_k0.tif`, `Gap_Total_t60_k5.tif`). These are GeoTIFFs in EPSG:4326 at ~1 km resolution, with values representing beds per 1,000 elderly (aged 60+). Auxiliary rasters include `Income_1km_2024.tif` (XGBoost-downscaled income), `Isochrone_Zones.tif` (OSRM-based travel time zones), and `Population_60plus_2025.tif` (elderly population projection).Administrative aggregation (`2_Administrative_Aggregation/`): CSV files at provincial, city, and county levels (e.g., `province_statistics.csv`, `county_statistics.csv`) contain `*_mean` (population-weighted mean accessibility) and `*_accessible_rate` (coverage rate, available only for baseline and constrained scenarios).Inequality metrics (`3_Inequality_Metrics/`): Files include `gini_national.csv`, `gini_provincial.csv`, `gini_city.csv`, and `gini_county_internal.csv`, providing Gini coefficients by bed type, threshold, and price sensitivity (k) across administrative levels.Facility points (`4_Facility_Points/`): `facility_points.csv` and `facility_points.shp` contain coordinates, bed capacity, prices, and nursing beds for over 42,000 facilities.Validation results (`5_Validation_Results/`): `Income_Validation/` includes `model_comparison_data.csv` and `feature_importance_data.csv`; `Travel_Time_Validation/` contains `raw_comparison_data.csv` (OSRM vs. commercial APIs); `Sensitivity_Analysis/` includes `TableS1_Key_k_Values.csv` and `TableS2_Marginal_Effects.csv`.3. Key value interpretationsItem Interpretation Raster values (`*_t*_k*.tif`) Units are beds per 1,000 elderly (aged 60+). A national benchmark of approximately 15 beds per 1,000 elderly can be used for reference. A value of 0 indicates no facility within the travel time threshold, while -9999 denotes NoData (unpopulated areas or water bodies). Coverage rate (`*_accessible_rate`) The proportion (0–1) of elderly within an administrative unit whose accessibility is ≥15 beds/1,000, available only for baseline and constrained scenarios. Gini coefficient (`gini` field in `gini_*.csv`) Ranges from 0 to 1; higher values indicate greater inequality in accessibility distribution across administrative units. Income raster (`Income_1km_2024.tif`) Per capita monthly disposable income (CNY/year), downscaled from county level using XGBoost. This data has a resolution of approximately 1 km and has been calibrated at the county, provincial, and city levels. Elderly population (`Population_60plus_2025.tif`) Projected number of individuals aged 60+ per pixel for 2025. Isochrone zones (`Isochrone_Zones.tif`) Integer codes 1–5 representing travel time from the nearest facility: 1=[0–15 min], 2=[15–30 min], 3=[30–60 min], 4=[60–90 min], 5=[>90 min]. 4. Recommended usageSpatial planning: `Total_t60_k0`Demand estimation: `Total_t60_k3`Equity assessment: `Gap_Total_t60_k3`Urban analysis: `Total_t30_k3`Rural analysis: `Total_t90_k3`Nursing care analysis: `Nursing_t60_k3`The 60-minute threshold with k=3 is recommended as the general default setting for policy-oriented analysis. 5. LimitationsTravel time is based on driving only.Affordability is simplified using income-based filtering.Facility prices are based on advertised fees and may not include all service costs.Community-based elderly care has not been adequately considered. Bed quality is not differentiated within each bed type.The dataset represents a cross-sectional snapshot rather than long-term dynamics.6. ContactXiao Wang ([e-mail])