oxigdal-analytics

Advanced geospatial analytics for enterprise workflows — part of the OxiGDAL ecosystem.

Features

• Hotspot Analysis: Getis-Ord Gi* and Moran's I spatial autocorrelation
• Spatial Clustering: K-means and DBSCAN for image classification and outlier detection
• Time Series Analysis: Trend detection (Mann-Kendall), anomaly detection, seasonal decomposition
• Change Detection: CVA (Change Vector Analysis), PCA-based, and multi-temporal analysis
• Interpolation: IDW (Inverse Distance Weighting) and Kriging for spatial surface generation
• Zonal Statistics: Weighted and multi-band zonal statistics by polygon zones
• Performance Profiling: Built-in timing and throughput metrics

Installation

[dependencies]
oxigdal-analytics = "0.1.3"

Usage

Hotspot Analysis (Getis-Ord Gi*)

use oxigdal_analytics::spatial::{getis_ord_gi_star, SpatialWeights};

let values = vec![1.0, 2.5, 8.0, 9.5, 2.0, 1.5];
let weights = SpatialWeights::queen_contiguity(&geometries)?;
let hotspots = getis_ord_gi_star(&values, &weights, 0.05)?;

for h in &hotspots {
    println!("Cell {}: z={:.3}, p={:.4}, cluster={:?}", h.id, h.z_score, h.p_value, h.cluster_type);
}

Spatial Clustering

use oxigdal_analytics::clustering::{KMeans, Dbscan};

// K-means for land cover classification
let kmeans = KMeans::new(5, 100)?;  // 5 classes, 100 iterations
let labels = kmeans.fit_predict(&pixel_features)?;

// DBSCAN for hotspot outlier detection
let dbscan = Dbscan::new(0.5, 3)?;  // eps=0.5, min_points=3
let clusters = dbscan.fit_predict(&point_features)?;

Time Series Trend Detection

use oxigdal_analytics::timeseries::{detect_trend, TrendMethod};

// Mann-Kendall trend test on NDVI time series
let trend = detect_trend(&ndvi_series, TrendMethod::MannKendall)?;
println!("Trend: {:?}, tau={:.3}, p={:.4}", trend.direction, trend.tau, trend.p_value);

Change Detection

use oxigdal_analytics::change::{detect_changes, ChangeMethod};

// Change Vector Analysis between two epochs
let changes = detect_changes(&before, &after, ChangeMethod::CVA)?;
println!("Changed pixels: {}/{}", changes.changed_count(), changes.total());

Interpolation

use oxigdal_analytics::interpolation::{idw, Kriging, KrigingModel};

// IDW interpolation from sparse sample points
let grid = idw(&sample_points, &sample_values, &target_grid, power: 2.0)?;

// Ordinary Kriging
let kriging = Kriging::new(KrigingModel::Spherical)?;
kriging.fit(&sample_points, &sample_values)?;
let surface = kriging.predict(&target_grid)?;

Feature Flags

COOLJAPAN Policies

• Pure Rust — no C/Fortran dependencies
• No unwrap() — all errors handled via Result<T, OxiError>
• Uses SciRS2-Core for scientific computing (not ndarray directly)
• Workspace dependencies via *.workspace = true

License

Apache-2.0 — Copyright (c) COOLJAPAN OU (Team Kitasan)