Correlating Resonance — Phase 3.1: Corpus Coding

What this does
Upload your finalised typology and the full corpus. Documents are coded in batches — each document independently scored against every concept in the typology. The DTM is exported automatically after each batch as a checkpoint. Review the DTM before proceeding to Phase 4 for threshold setting, binarisation, and cluster analysis. If the coding looks wrong, adjust the configuration or return to Phase 2 to revise the typology.

Configuration

Coding mode

Rating (0–3)

Frequency count

0Absent 1Mentioned — brief reference only 2Important — discussed and emphasised 3Central — core focus of document

Processing settings

Batch size

Documents processed before each automatic DTM checkpoint export.

Text extract limit (characters)

Characters extracted per document. Higher = more context, more tokens.

Vision fallback threshold (characters)

If extracted text is below this length the PDF is sent directly to the API for vision-based coding.

Typology CSV

📋

Drop CSV here or click to browse

Finalised typology from Phase 2

Corpus documents

📄

Select multiple files Drop PDF or TXT files here, or click to browse and select all corpus documents at once

.pdf and .txt · select all at once using Shift+click or Ctrl+click

📁 Select
folder

Resume from checkpoint — optional

⏸

Drop a checkpoint DTM here or click to browse

Documents already in this file will be skipped

Load typology and corpus to begin

Coding progress

0 / 0

Document-Term Matrix

Cells highlighted in amber indicate low-confidence scores flagged by the LLM. Proceed to Phase 3.2 for threshold setting and binarisation.