planx_geostats

Changelog

[0.9.23] - 2026-08-08

• Update documentation link to GEOSTATS_REFERENCE_MANUAL.html

[0.9.22] - 2026-08-07

• Added online user manual link (https://yusufeminoglu.github.io/planx_geostats/GEOSTATS_REFERENCE_MANUAL.html) and GitHub repository star call-to-action.

[0.9.21] - 2026-08-07

• Add floating Save as PDF button to reference manual

[0.9.20] - 2026-08-07

• 2.5x manual expansion: 300+ equations, 215+ DOI refs, HelpUrlMixin, GitHub Pages

[0.9.19] - 2026-07-14

• Add Bivariate Local Moran’s I (Bivariate LISA) algorithm and unique icon

[0.9.18] - 2026-06-18

• docs: add CITATION.cff for Zenodo DOI integration

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.9.17] - 2026-06-05

• Added Spatial Inequality (Gini and Spatial Gini) to the Urban Pattern Scan group.
• Added classic Gini, neighbor/non-neighbor Gini components, spatial Gini share, spatial polarization, and permutation inference.
• Added optional CSV/JSON exports, HTML analyst guidance, algorithm icon, provider registration, smoke coverage, and release-gate count alignment.
• Added a full QGIS runtime algorithm matrix that executes every GeoStats Processing algorithm against bundled sample data on QGIS 3 LTR and QGIS 4.
• Restored the root plugin icon path by packaging icons/icon.png and pointing metadata.txt to the PNG asset used by the Processing provider.
• Fixed Similarity Search expression-context handling and geometry-type checks in OLS/GLR residual diagnostics.
• Hardened optional PySAL/MGWR dependency diagnostics so broken import-time dependency stacks are reported with the exact installer command.

[0.9.15] - 2026-05-30

• Revised plugin icon to SVG

[0.9.14] - 2026-05-27

• Code quality and Hub submission hygiene.
  • Added setup.cfg (flake8: W503, E203 disabled; max-line-length=120).
  • Fixed 21 flake8 issues across 15 files: unused QVariant imports, missing blank lines, continuation indentation hoisted from f-strings, trailing whitespace stripped.
  • All 42 Python files compile cleanly. Zero non-E501 flake8 issues.

[0.9.13] - 2026-05-26

• GeoStats provider/runtime smoke coverage updates and QGIS 3.40+/4 compatibility validation.

[0.9.12] - 2026-05-26

• Maintenance release: refreshed Plugin Hub package after QGIS 3 and QGIS 4 compatibility validation.

[0.9.10] - 2026-05-22

Changed

• Completed beta stabilization for the PlanX GeoStats Lab Processing provider.
• Centralized optional dependency error guidance for MGWR, Spatial Lag Regression, and Spatial Error Regression so missing-library failures point users to GeoStats Library Status and the explicit installer workflow.
• Hardened output-layer metadata handling so field descriptions are retained as custom layer properties even when a provider cannot apply field aliases.
• Hardened centroid and geometry handling across distance, center, distribution, pattern, GWR, MGWR, and export workflows so missing, empty, invalid, or provider-problem geometries are skipped safely instead of crashing the algorithm.

Added

• Added QGIS-independent smoke coverage for dependency guidance, metadata resilience, null/empty geometry handling, and centroid failure handling.
• Added release-gate coverage to keep core decision helpers, output metadata contracts, manual QA documentation, and package verification aligned.

[0.9.9] - 2026-05-22

Added

• Added personalized Workflow Advisor inputs for analysis goal, geometry context, outcome type, and explanatory-variable availability.
• Added a personalized recommendation section that returns a suggested tool sequence and pre-trust checks.

[0.9.8] - 2026-05-22

Added

• Added GeoStats Workflow Advisor, a Processing report that maps planning questions to recommended GeoStats tools, prerequisites, outputs, and follow-up checks.
• Added method assumptions, common pitfalls, safer moves, and bundled-sample starter recipes to the Workflow Advisor report.
• Added smoke coverage for Workflow Advisor guidance sections and a minimum provider algorithm-count guard.
• Added a manual QA matrix covering setup, pattern analysis, geometry summaries, modeling workflows, symbology, reports, and release gates.
• Added shared report-guidance and output-layer metadata helpers for more consistent professional reports and layer outputs.
• Added VIF diagnostics to regression quality reporting and rank/score auditing to Model Comparison Matrix.

[0.9.7] - 2026-05-22

Added

• Added smoke coverage that keeps metadata.txt, CHANGELOG.md, and the README release verification command synchronized on the same plugin version.

[0.9.6] - 2026-05-22

Changed

• Expanded the Sample Dataset Guide QA fixture table to list every model-output layer explicitly instead of using a wildcard label.

Added

• Added smoke coverage that verifies the generated Sample Dataset Guide source mentions every loadable sample and QA layer.

[0.9.5] - 2026-05-22

Added

• Added smoke coverage that keeps the Sample Dataset Guide load options synchronized with the bundled synthetic QA GeoPackage layers.
• Added an explicit loading-modes section to the Sample Dataset Guide HTML report.

[0.9.4] - 2026-05-22

Added

• Added planx_geostats_synthetic_qa.gpkg, a separate deterministic QA fixture with point, line, polygon, and minimal model-output layers.
• Added synthetic QA sample smoke checks for layer presence, geometry types, feature counts, model-output fields, binary fields, and count fields.
• Added synthetic QA core-engine smoke checks for ANN, Ripley’s K, GLR logistic/Poisson, and Linear Directional Mean.
• Added static smoke coverage for multipart guards around direct asPolyline() / asPolygon() geometry conversions.
• Added a Sample Dataset Guide load selector for the Izmir planning sample, the synthetic QA fixture, or both datasets.

Changed

• Expanded the Sample Dataset Guide and sample-data README to document both the Izmir planning sample and the synthetic QA fixture.
• Removed several unused runtime imports detected during deep QA cleanup.
• Hardened release code paths so the packaged plugin scans with no Bandit findings in local validation.
• Updated release validation workflow for version 0.9.4.

[0.9.3] - 2026-05-21

Fixed

• Fixed Incremental Spatial Autocorrelation HTML report generation on QGIS 3.40 by avoiding html module shadowing.
• Fixed KNN spatial weight generation on QGIS 3.40 by using the current QgsSpatialIndex.nearestNeighbor API with a legacy fallback.

Added

• Added smoke coverage for QGIS spatial-index API compatibility and HTML module shadowing in report writers.

[0.9.2] - 2026-05-21

Changed

• Removed developer-only tests/ files from release zip packages so QGIS Hub security scanning only evaluates runtime plugin code.
• Removed hidden .gitignore metadata from release zip packages; it remains in the source repository only.
• Added release-zip verification to the developer validation workflow.

Fixed

• Resolved the QGIS Hub Bandit block caused by source-only SQLite smoke-test SQL strings being included in the uploaded plugin zip.
• Cleaned several low-risk Hub quality warnings, including unused imports, import ordering, unused local variables, and whitespace style issues.
• Replaced a runtime assert in the library installer with an explicit Processing exception.

[0.9.1] - 2026-05-21

Added

• Added a QGIS-independent provider catalog smoke test to protect algorithm registration, unique Processing ids, display names, and workflow group coverage.
• Expanded core smoke coverage for GLR family validation and Poisson likelihood accounting.
• Added a plugin-local .gitignore for Python caches, IDE files, OS files, and QGIS backup artifacts.
• Added simple, distinct PNG icons for every Processing algorithm and wired each algorithm to its own icon.

Fixed

• Corrected Poisson GLR log-likelihood and AIC calculation to include the count factorial term.

Changed

• Kept optional-library status and installation workflows only under 00 | Setup and Diagnostics; removed the separate GeoStats Libraries menu/toolbar UI.

Removed

• Removed the unused dependency dialog UI now that setup workflows are Processing-only.

[0.9.0] - 2026-05-20

Changed

• Renamed the Processing provider to PlanX GeoStats Lab.
• Reorganized tools into English PlanX planning workflow groups: Data Preparation and Neighborhoods, Urban Pattern Scan, Hot Spots and Spatial Outliers, Centers Direction and Dispersion, and Models and Scenarios.
• Moved dependency installation out of Processing and into the built-in GeoStats Libraries helper under the PlanX GeoStats Lab menu.
• Expanded the GeoStats Libraries helper with detailed English guide text, package role explanations, command preview guidance, install-mode guidance, and restart guidance.
• Moved the GeoStats Libraries menu action under PlanX GeoStats Lab > GeoStats Libraries so it is scoped to the GeoStats plugin rather than the general PlanX menu.
• Added a GeoStats Libraries toolbar action so the helper remains visible even when QGIS nests plugin menus differently.
• Corrected dependency command generation so QGIS application executables are not shown as pip runners; the helper and status report now distinguish the QGIS host application from the Python executable used for pip.
• Expanded the library status report with a clear “How to install” section for guided installation and manual OSGeo Shell usage.
• Added a Processing Toolbox installer, Install / Update GeoStats Libraries, under 00 | Setup and Diagnostics for profiles where the menu helper is not visible.
• Improved the Processing installer with preview-only behavior, detected path logging, and smarter default mode selection.
• Fixed Global Moran’s I z-score and p-value calculation in the shared statistics engine.
• Fixed General G value alignment for QGIS layers whose feature IDs are not contiguous zero-based indices.
• Added neighbor-count diagnostic fields to Gi* and Local Moran output layers.
• Added shared analysis diagnostics for numeric quality, CRS warnings, and neighborhood graph summaries.
• Upgraded Global Moran and General G reports with executive summaries, diagnostics, caveats, and recommended next actions.
• Added OLS model-quality diagnostics for sample size, skipped records, near-constant predictors, multicollinearity, and condition number.
• Enhanced Exploratory Regression reports with model-quality checks, AICc rank reasons, and final-model review guidance.
• Upgraded Average Nearest Neighbor and Incremental Autocorrelation reports with executive summaries, CRS caveats, and next-action guidance.
• Added neighbor-support diagnostics to Incremental Autocorrelation distance scan results.
• Upgraded Sensitivity Test reports with Monte Carlo interpretation, input diagnostics, neighborhood diagnostics, caveats, and next-action guidance.
• Upgraded GWR outputs and reports with local support diagnostics, model-quality checks, CRS caveats, and bandwidth interpretation guidance.
• Added audit fields to Multivariate Clustering and Similarity Search outputs for cluster size, cluster distance, similarity percentile, and similarity tier.
• Added weight-quality checks and audit fields to Mean Center, Standard Distance, and Standard Deviational Ellipse outputs.
• Added Ripley’s K-Function as a new urban pattern scan tool with L(d)-d diagnostics, neighborhood support, and planning caveats.
• Added Generalized Linear Regression (GLR) with Gaussian, Logistic, and Poisson model families.
• Added Spatial Autoregression as a PySAL spreg-based spatial lag model with rho diagnostics, neighbor support, output audit fields, and a PlanX analyst report.
• Added Spatial Error Regression (SEM) with lambda diagnostics, residual spatial checks, audit fields, and model interpretation guidance.
• Added Multiscale Geographically Weighted Regression (MGWR) using PySAL mgwr with variable-specific bandwidths, local coefficient audit fields, and a scale-focused analyst report.
• Added Model Comparison Matrix for comparing PlanX model output layers by fit metrics, coverage, bias, and residual spatial autocorrelation.
• Added curated English-schema Izmir neighborhood sample GeoPackage for development, demos, manual QA, and regression workflow testing.
• Added Sample Dataset Guide to load the bundled sample layer from the Processing Toolbox and explain recommended workflows.
• Added Data Readiness Audit under setup diagnostics to review CRS risk, numeric field completeness, constant indicators, sample workflow readiness, and recommended next actions before formal analysis.
• Expanded Data Readiness Audit with geometry QA for empty, invalid, and multipart features before contiguity, local statistics, and distance-based workflows.
• Added a multicollinearity screen to Data Readiness Audit so high-correlation numeric field pairs are flagged before OLS, GLR, GWR, MGWR, and spatial regression workflows.
• Added analysis-role suggestions to Data Readiness Audit so numeric fields are labeled as target/pattern variables, explanatory candidates, count/intensity indicators, or review-only fields with likely tool guidance.
• Added optional field-level CSV export to Data Readiness Audit for spreadsheet QA logs and audit handoffs.
• Expanded Data Readiness Audit workflow guidance with target fields, candidate explanatory fields, recommended tool sequences, and planning-purpose notes for starter analyses.
• Added optional full JSON export to Data Readiness Audit for reproducible audit handoffs, automation, and downstream QA integration.
• Added distribution-shape diagnostics to Data Readiness Audit, including median, quartiles, IQR, skewness, IQR outlier counts, distribution notes, and risk warnings for skewed or outlier-heavy fields.
• Expanded bundled sample-data smoke tests to enforce English schema stability, numeric field types, expected value ranges, complete critical fields, and sufficient variation for analysis workflows.
• Added Bivariate Spatial Association (Lee’s L) for local cross-variable neighborhood diagnostics.
• Added shared residual spatial autocorrelation diagnostics to GLR, GWR, Spatial Autoregression, and MGWR reports.
• Added QGIS-independent core smoke tests for the main statistics engines and diagnostics helpers.
• Documented the smoke-test and packaging validation workflow in the README.
• Cleaned remaining special symbols from Processing feedback and generated reports.
• Replaced remaining legacy hyphenated labels in logs and generated HTML reports with PlanX GeoStats Lab.
• Reworked optional type hints to avoid Python 3.10-only union syntax in QGIS 3.28-era Python environments.
• Added an Exploratory Regression safety limit so overly large candidate-variable searches fail fast with guidance instead of tying up QGIS.

Added

• Registered Central Feature and Incremental Spatial Autocorrelation in the provider.
• Exploratory Regression tool for screening OLS variable combinations and ranking candidate models by AICc.
• Exploratory Regression report now records how many candidate models were estimated.
• GeoStats Library Status diagnostic algorithm under 00 | Setup and Diagnostics, producing a non-installing HTML dependency report and QGIS Python command preview.
• Install / Update GeoStats Libraries diagnostic algorithm with explicit approval, command logging, OSGeo Shell mode, and restart guidance.

[0.8.0] - 2026-05-20

Added

• Assessing Sensitivity toolset:
  • Attribute Randomization Sensitivity Test — Monte Carlo permutation simulation for Global Moran’s I with SVG histogram HTML report.
• Measuring Geographic Distributions toolset:
  • Linear Directional Mean — circular weighted mean orientation for line features with trend line output.

[0.7.0] - 2026-05-20

Added

• Mapping Clusters toolset:
  • Multivariate Clustering (K-Means) — K-Means++ initialization, Z-score standardization, auto categorized symbology.
• Utilities toolset:
  • Export Feature Attributes to CSV/ASCII — configurable delimiter, optional centroid coordinates.

[0.6.0] - 2026-05-20

Added

• Mapping Clusters toolset:
  • Similarity Search — Z-score attribute profiling with Euclidean/Manhattan distance metrics, auto graduated symbology.
• Spatial Component Utilities toolset:
  • Calculate Distance Band from Neighbor Count — k-th neighbor distance statistics with percentile HTML report.

[0.5.0] - 2026-05-20

Added

• Measuring Geographic Distributions toolset:
  • Median Center — Weiszfeld’s algorithm for weighted spatial median.
• Analyzing Patterns toolset:
  • High/Low Clustering (Getis-Ord General G) — global G index with randomization variance, HTML report.

[0.4.0] - 2026-05-20

Added

• Analyzing Patterns toolset:
  • Average Nearest Neighbor (ANN) — chunked KDTree with z-score significance testing, HTML report.
• Measuring Geographic Distributions toolset:
  • Standard Distance — weighted circular dispersion polygon at 1/2/3 std dev.
• Modeling Spatial Relationships toolset:
  • Geographically Weighted Regression (GWR) — Fixed Gaussian, Fixed/Adaptive Bisquare kernels, local coefficients, auto graduated symbology on local R².

[0.3.0] - 2026-05-20

Added

• Analyzing Patterns toolset:
  • Global Moran’s I — spatial autocorrelation with randomization variance.
• Modeling Spatial Relationships toolset:
  • OLS Spatial Regression — ordinary least squares with diagnostics HTML report.

[0.2.0] - 2026-05-20

Added

• Mapping Clusters toolset:
  • Local Moran’s I (LISA) — local spatial autocorrelation with HH/HL/LH/LL classification, auto categorized symbology.
• Measuring Geographic Distributions toolset:
  • Mean Center — weighted arithmetic mean center.
  • Standard Deviational Ellipse (SDE) — orientation, semi-axes, ellipse polygon output.
• Dependency Installer utility algorithm.

[0.1.0] - 2026-05-20

Added

• Mapping Clusters toolset:
  • Getis-Ord Gi* Hot Spot Analysis — z-score and p-value with auto graduated symbology.
• Initial plugin skeleton, metadata, icons, and Processing provider registration.

This site is open source. Improve this page.