Bring the profile
into focus.

Start with a measured CPT, inspect the intervals, then retrieve observed precedent fingerprints. Rankings are evidence for review—not calibrated probability or an engineering conclusion.

FINGERPRINTS5,668unique physical CPTs

EVENT RECORDS17,004attached observations

QUERYRAWawaiting validation

Add a CPTCSV, TSV, or a pasted measurement table

Choose CSV / TSV

OR PASTE A TABLE

CONTEXT & METADATADISPLAY NAME

EARTHQUAKE EVENT required for experimental context projectionNEIGHBOURS

PGA (g) · 0–3Magnitude Mw · 4–9.5 · audit onlyGWT depth (m) · 0–30Unit weight (kN/m³) · 10–30 · audit only

Validation and resampling happen only in the local service. Raw source rows are not added to the research corpus.

Profile fingerprintAdd a table to begin

Your CPT fingerprint appears here.

Map depth and qc, then preview the actual measured profile before retrieval.

Compare two precedent lensesShape alone, then an experimental context adjustment

PURE FINGERPRINTPROFILE RMSE v2

Depth-indexed resemblance with continuous overlap support.

EVENT & OUTCOMES EXCLUDED FROM RANK

EXPERIMENTAL CONTEXTPROFILE PROJECTION

75% standardized profile shape + 25% supported scalar context for the selected event.

ATTEMPTED: qc · fs · Ic · Fr · depth · PGA
Mw & unit weight: audit only · not frozen C4

Neither list is a probability or risk result. Outcomes are attached only after rank and never influence either ordering.

04A

Pure fingerprint matchesValidate a profile, then retrieve

Pure fingerprint matches appear here.

Validate a profile to compare its measured geometry.

04B

Experimental context projectionRequires event, PGA, GWT, and fs support

Context-adjusted precedents appear here.

Load the complete demo or supply event, PGA, measurement GWT, and fs.

KNOWN CASE DEMO

The original known-case flow remains available as a controlled corpus demonstration. External CPTs are the primary workspace; browser code never calculates an approximation or downloads the private corpus.

PROFILE INTERPRETATION

Coloured bands are Ic-derived soil-behaviour zones, not geology or lithology. Gaps and invalid intervals are evidence of missing support, not material differences.

OUTCOME INTERPRETATION

A neighbour outcome fraction is retrieval evidence, not calibrated probability, a risk estimate, or engineering judgment.

Bring the profileinto focus.

Bring the profile
into focus.