Orebit Assay

Stage 02 · Orebit Assay · Statistics, validation, compositing for Orebit Resource

KCMI 2017 compliant JORC 2012 cross-ref 12 tab workflow data not loaded

Dashboard

GeoSuite Assay · composite preparation for JORC/KCMI Table 1

Composites

Elements

Domains

Lithologies

Goal: Prepare a drillhole composite dataset that meets JORC/KCMI Table 1 Section 1, ready to feed the Orebit Resource (Artefact B).

Standards: JORC Code 2012, KCMI 2017, NI 43-101, industry best-practice.

Statistics summary

—

Domain summary

—

Workflow guide & reference — 11 steps, decision criteria, glossary

11-Step Workflow

#	Step	Tab	Main Output	Decision Gate
1	Import Data	Upload	Parsed CSV + column mapping	Auto-detect ≥80% critical columns
2	Visual Inspection	Data	Welllog + sample table	No obvious depth/grade artifacts
3	Bivariate EDA	Bivariate	Correlation heatmap + scatter matrix	Multi-element relationships known
4	Multivariate EDA	Multivariate	PCA biplot + domain boxplots	PC1/PC2 variance explained ≥ 60%
5	Spacing	Spacing	NN distance distribution + density map	Median spacing <50% est. range
6	Compositing	Composite	1m weighted-average composites	CV < raw, n_samples preserved
7	EDA Statistics	Stats	n / mean / median / SD / CV / P10–P99	CV known, log-norm checked
8	Top-Cut	Top-Cut	Threshold + impact (% samples, metal loss)	Cut affects <1% (P99) or <5% (P95)
9	Grade by Litho	Litho	Per-litho stats + boxplot ranking	Host vs barren litho identified
10	Domain Modeling	Domain	Combined geo+grade domains	ANOVA p<0.05 between domains
11	Report + KCMI	Report	Before/after summary + KCMI Table 1	All gates passed, CPI sign-off

Decision Criteria per Stage

Compositing length

1m default — porphyry Cu, close-spaced vein Au, mass disseminated
2m — mid-grade epithermal, low-grade Ni laterite
0.5m — narrow-vein high-grade Au, Ag

Rule of thumb: composite length ≈ ½ × selective mining unit (SMU), or ≈ average assay interval.

Top-cut threshold

P99 default — VMS Cu-Zn, porphyry Cu-Au (low CV < 1.5)
P98 — disseminated Au, mid-CV (1.5–3)
P95 — high-nugget Au vein, high CV (> 3)
No cut — coal, iron ore, bulk commodities

Spacing → Classification (KCMI 2017 Table 1 Section 5)

Measured — spacing < ½ variogram range, or two consistent data sources
Indicated — spacing ≈ variogram range
Inferred — spacing > variogram range, continuity still confirmed

Note: this tool gives a first-pass estimate using NN. Final classification requires kriging variance.

Domain validation (ANOVA)

p < 0.01 — domains highly distinct ✓
0.01 ≤ p < 0.05 — domains acceptable ✓
p ≥ 0.05 — domains not significantly different → redefine

Suitable for

Drilling completed or closing interval → ready for resource estimation
At least 30 holes + 1,000 assay samples
Compositing not yet done, or needs re-compositing
Preparing JORC/KCMI Table 1 Section 1 (Sampling Techniques and Data)
First-pass EDA for CPI / senior geologist briefing

Not suitable for

Greenfield (data < 30 holes) → use Stage 01 Targeting Suite
Pre-collar / scout drilling → this is for close-spaced infill
Replacement for commercial estimation software
Final KCMI report without CPI sign-off

Technical Glossary

CV (Coefficient of Variation)	SD ÷ mean. CV > 1.5 = high variance, top-cut needed. CV > 3 = critical, possibly multi-population.
Top-cut / Capping	Capping extreme values at a percentile (commonly P99) to reduce outlier impact on estimates.
Compositing	Combining assay intervals into uniform lengths weighted by original interval length (length-weighted average).
Domain	Sub-population with distinct statistical character (mean, variance, spatial continuity) → must be estimated separately.
ANOVA	Statistical test for whether means differ significantly between groups. High F-statistic + p < 0.05 = distinct groups.
NN distance	Nearest Neighbor distance — distance from one collar to the closest collar. Median NN ≈ effective drill spacing.
P99 / P95	99th / 95th percentile — value below which 99% / 95% of samples fall.
Welllog	Visualization of grade vs depth for a single hole, typically color-coded by lithology.
KCMI Table 1	Section 1 (Sampling Techniques and Data) from KCMI 2017 — required for public resource/reserve reports in Indonesia.
CPI	Competent Person Indonesia — geologist registered with PERHAPI or IAGI authorized to sign off KCMI reports.

Get started with the sample data → click the "Data" tab to inspect, or "Upload" to import your own.

Analysis Scope

Stage 02 — GeoSuite Assay

Prepare a drillhole composite dataset that meets JORC/KCMI Table 1 Section 1, ready for GeoSuite Resource.

✓ What this does:

Length-weighted compositing (0.5–10 m)
Domain assignment (manual + K-Means)
Histogram, boxplot, CV, top-cut matrix
EDA Report PDF + KCMI Table 1 draft

Not included:

Cell declustering → SGeMS / R gstat
Auto top-cut suggestion (matrix only)
Variogram cloud → Stage 03

Output: domain-tagged composite CSV + EDA Report PDF + KCMI narrative

Quick Start Tutorial

Upload — Import master CSV from Stage 01 (or use sample)
Composite — Pick an interval (default 1m)
Stats & Top-Cut — Review CV, set top-cut threshold
Domain — Assign by lithology or auto-cluster
Export — Download composite CSV for Stage 03

Tip: Start with the Tour button in the top-right header for a guided walkthrough.

Sample Dataset Info

Upload Data — Import CSV

Format: CSV with header row. Columns are auto-detected (case-insensitive: hole_id, HoleID, BHID all recognized).

Multi-file: Upload collar.csv + assay.csv + litho.csv together; the system auto-merges by hole_id.

Privacy: All processing runs client-side. Data is never sent to a server.

Click or drag & drop a CSV file here

Multi-file support: collar.csv + assay.csv + litho.csv

— or —

Recognized Column Formats

Field	Aliases (auto-detected)	Required?
hole_id	HoleID, BHID, DHID, HOLE, hole	✓
from_m	FROM, from, From, depth_from, fr	✓
to_m	TO, to, To, depth_to	✓
au_gpt	AU, Au, au_ppm, AU_GPT, gold	✓ (or primary commodity)
cu_pct	CU, Cu, cu_ppm, copper	optional
lithology	LITH, lith, rock_type, geology	optional (warn if missing)
midx, midy, midz	x, y, z, easting, northing, rl, elevation	optional (for spatial plots)
density, recovery_pct, rqd_pct	SG, density_t/m3, rec, RQD	optional (QA/QC)

Reset

Restore the embedded sample data (1,241 composites · 40 holes, Thalanga RGMWP).

Data — Welllog & Samples

Embedded dataset: 1,241 composites · 40 holes · 16 columns · Thalanga VMS (RGMWP prospect, output of Stage 01).

Welllog — Multi-element Trace

Per-hole multi-element subplot. Color-coded bars = grade; side-by-side bars compare mineralization zoning across elements. Litho color overlay sits to the left of each track.

Hole: Elements: Log scale x-axis

Data Table

Bivariate Analysis — Correlation & Scatter Matrix

Correlation heatmap + pairwise scatter plots to identify relationships between numeric variables. High correlation → redundancy or mineral assemblage; low correlation → independent processes.

Multivariate Analysis — PCA & Domain Boxplots

Principal Component Analysis (PCA) reduces dimensionality. PC1 is usually grade-dominated; PC2 can reveal secondary processes (alteration, litho control). Boxplot by domain validates domain definitions.

Spacing & Density Analysis

Industry rule: median spacing ≤ 50% of variogram range. Drill density drives the resource classification level (Measured / Indicated / Inferred).

Compositing — Weighted by Length

Compositing standardizes assay intervals to a fixed length. Method: length-weighted average. Default 1m (matches porphyry/disseminated SMU).

EDA Stats — Comprehensive Distribution Analysis

High CV (>1.5) flags the need for top-cut. Log-skewed distributions → candidates for log-transform before kriging.

Auto-Pilot

Automatically runs P99 top-cut, K-Means clustering, and ANOVA validation.

Top-Cut Analysis — Percentile Method

P99 = standard for VMS/porphyry (affects <1% of samples). P95 for high-nugget Au.

Grade by Lithology

Per-lithology statistics → identify rock types hosting mineralization. Prerequisite for geological domain modeling.

Domain Modeling — Combined Geological + Grade

Domain MUST be defined by geological boundaries + grade continuity, NOT grade shells alone.

Report & KCMI Table 1

Export Options

Multi-page PDF with summary + visualizations + KCMI checklist.

CPI Disclaimer

This output is not a final KCMI Table 1. The final commentary must be reviewed and signed off by a CPI registered with PERHAPI/IAGI before official publication.

KCMI Table 1 — Auto-fill Commentary

Section 1: Sampling Techniques and Data

Data Preview

First 10 rows of the composite dataset that will be exported to Stage 03 (Resource).

Export History

Audit trail of every download this session (max 20). Persists in browser localStorage.