running-kriging-compute

Running Kriging Compute

This example demonstrates a complete geostatistical workflow using the Evo Python SDK:

1. Load downhole assay data as a PointSet
2. Define a variogram model for spatial correlation
3. Visualize the pointset and variogram together with Plotly
4. Run kriging estimation using Evo Compute (WIP)

Overview

The workflow uses the WP_assay.csv dataset containing copper (CU_pct) and gold (AU_gpt) assay values from 55 downhole. We'll:

• Create a PointSet from the CSV data
• Define a nested spherical Variogram model for copper grades
• Extract and scale Ellipsoid objects for search neighborhoods
• Visualize the data, variogram curves, and anisotropy ellipsoids with Plotly
• Set up kriging estimation parameters (WIP)

Dataset Characteristics

• 8,332 sample points from 55 downhole
• Spatial extent: ~936m (X) × ~1,416m (Y) × ~855m (Z)
• Coordinate system: EPSG:32650 (UTM Zone 50N)
• Target attribute: CU_pct (copper percentage)
  • Mean: 0.95%, Variance: 0.84

Variogram Model

The variogram uses two nested spherical structures aligned with the dominant orientation of the downhole data:

• Nugget: 0.08 (~10% nugget effect)
• Short-range structure: Contribution 0.25, ranges 80m × 60m × 40m
• Long-range structure: Contribution 0.51, ranges 250m × 180m × 100m
• Anisotropy: Dip 70°, Azimuth 15° (NNE strike direction)

Kriging Compute

The notebook includes work-in-progress sections demonstrating:

• Creating a target BlockModel for estimation
• Configuring KrigingParameters with search neighborhoods
• Running kriging tasks with evo.compute
• Running multiple scenarios in parallel for sensitivity analysis

Requirements

• Python 3.10+
• Seequent account with Evo entitlement
• Evo application credentials (client ID and redirect URL)

Quick Start

1. Open running-kriging-compute.ipynb in Jupyter
2. Update the client_id and redirect_url with your Evo app credentials
3. Run the cells to create objects and visualize the geostatistical model