Before you begin
This phase expects a reviewed and corrected DTM CSV from Phase 3.1. Low-confidence cells (marked [?]) and any anomalous scores should be resolved in the CSV before uploading. The analysis proceeds in four steps: binarisation → TTM construction → cluster analysis → frame naming. Download the sorted TTM after Step 4.2 and compare it against the clustered output from Step 4.3 to validate the cluster structure before proceeding to naming.
Loading Python environment — this takes 15–30 seconds on first load…
Reviewed DTM
📊
Drop reviewed DTM CSV here or click to browse
Output from Phase 3.1 with low-confidence cells resolved
Each pixel-column is a concept; each pixel-row is a document. Brighter cells indicate higher scores. The green/grey boundary shifts as you adjust the threshold.