{"categories":[{"label":"Scientific/Engineering","url":"https://skillfed.io/packages/category/scientific-engineering/8"},{"label":"GIS","url":"https://skillfed.io/packages/category/scientific-engineering-gis"}],"enrichment":{"capability":"Calibrates multiscale and traditional geographically weighted regression (GWR/MGWR) models to analyze spatial heterogeneity in regression relationships across geographic regions.","skillfed_tags":["spatial-statistics","geospatial-analysis","regression-modeling"],"use_cases":["Analyze how housing price determinants vary geographically\u2014e.g., proximity to transit matters more in urban areas than rural ones.","Model disease incidence or health outcomes across regions where risk factors operate at different spatial scales.","Estimate local effects of environmental variables (pollution, temperature) on ecological or agricultural outcomes.","Conduct spatial heterogeneity tests to determine whether a global regression model is appropriate or if local variation is significant.","Generate spatially-varying predictions and confidence intervals for regression surfaces across a study region."],"what_it_does":"MGWR is a Python implementation of multiscale geographically weighted regression, a statistical method for modeling spatial relationships where regression coefficients vary across geographic locations. It extends traditional GWR by allowing different spatial scales (bandwidths) for different covariates, capturing the idea that some processes operate at finer or coarser geographic resolutions than others.\n\nThe package handles model calibration for Gaussian, Poisson, and binomial probability distributions; bandwidth selection via golden section or equal interval search; spatial prediction; and model diagnostics including multiple hypothesis test correction and local collinearity assessment. It builds on sparse generalized linear modeling (spglm) and integrates with the PySAL ecosystem (libpysal, spreg). Parallel computing is available for computationally intensive GWR and MGWR workflows.","worth_installing":"Yes. The package is production-stable, actively maintained, has no known vulnerabilities, and low installation friction. It fills a specialized but well-established niche in spatial statistics. Install it if you need to model geographic variation in regression relationships; skip it if your analysis is non-spatial or you need only global regression coefficients."},"id":"mgwr","links":{"html":"https://skillfed.io/packages/mgwr","md":"https://skillfed.io/packages/mgwr.md","pypi":"https://pypi.org/project/mgwr/"},"maintenance":{"status":"active"},"meta":{"latest_release":"2024-01-06","license_spdx":null,"license_treatment":"permissive","name":"mgwr","python_support":"supports_current","summary":"multiscale geographically weighted regression"},"popularity":{"monthly_downloads":101811,"position":12914,"tier":"top_15000"},"security":{"n_vulnerabilities":0},"version":"2.2.1"}