{"name": "Downscale Anything with Random Forest", "package_name": "downscale_anything_rf_fast", "description": "Downscale Anything with Random Forest is a QGIS Processing plugin that allows\r\nflexible dasymetric downscaling of coarse rasters to user-defined resolutions.\r\nThe algorithm uses high-resolution covariates and Random Forest regression to redistribute values.\r\nKey features include:\r\n1. User-defined target resolution.\r\n2. Random Forest regression with fast pixel sampling.\r\n3. Support for random or spatial-block splitting.\r\n4. Evaluation metrics (R\u00b2, RMSE, MAE) and OOB score.\r\n5. Outputs raster, CSV (evaluation + feature importance), and PNG plots.", "about": "This plugin provides a practical way to perform machine-learning-based downscaling in QGIS.\r\nIt is suitable for spatial analysis, environmental modeling, and urban/regional studies.\r\nExample use cases include downscaling VIIRS Nighttime Lights (avg_rad), Land Surface Temperature (LST),\r\npopulation density (e.g., WorldPop), and other gridded indicators (e.g., PM2.5, precipitation,\r\nsocio-economic indices).\r\nCreated by Firman Afrianto.", "homepage": "https://github.com/firmanaf/DownscaleAnythingRF", "repository": "https://github.com/firmanaf/DownscaleAnythingRF", "tracker": "https://github.com/firmanaf/DownscaleAnythingRF/issues", "author": "Firman Afrianto", "tags": ["raster", "regression", "spatial analysis", "random forest", "urban planning", "ai", "downscaling"], "downloads": 1095, "latest_version": "1.0.0", "versions": [{"version": "1.0.0", "experimental": false, "qgis_min": "3.0.0", "qgis_max": "3.99.0", "downloads": 1109, "uploaded_by": "firmanafrianto", "upload_datetime": "2025-08-20T11:38:43.012008"}]}