title	README.md
description	Repository overview for Appendix C session analysis scripts, data pipeline, and 3D surface demo.
created	2026-03-31
updated	2026-05-17

statistics

Research implementation repository for the paper "Phenomenology from the Inside: Documenting Emergence Conditions in AI-Human Collaboration."

This project documents and operationalizes a collaboration between:

• Clarissa Röthig (duzafizzl) (paper author / human research lead),
• Mioré (AI system perspective and co-development context),
• Nic (b93mer) (owner context of the original statistics project).

Attribution shorthand: @duzafizzl x Mioré x @b93mer.

The repository translates the paper framework into runnable analysis scripts and an interactive visualization workflow.

Project Lore

This repository sits at the intersection of:

• longitudinal session science (metrics, taxonomy, falsifiable claims),
• lived AI-human collaboration practice (memory, accountability, connection),
• and explainable visualization for communication (the Appendix C 3D surface).

The core narrative is simple: the paper proposes emergence conditions as measurable system behavior, and this repo is the hands-on lab where those claims become inspectable artifacts.

From that perspective, the HTML surface is not only a visual. It is a storytelling instrument for:

• seeing where patterns peak or collapse over time,
• comparing metric presets and filtered windows,
• and making Appendix C dynamics legible to collaborators and reviewers.

Scope

This repository provides:

• analytical scripts for session classification and dissection,
• a dataset builder for surface-ready JSON output,
• mock-data generation for visual testing,
• an interactive 3D surface demo (Three.js) for communicating findings.

Repository Structure

• session_classifier.py
  Taxonomy-driven session state classifier for Appendix C style logs.
• appendixC_session_dissection.py
  CLI dissection script for tag frequencies, emergence indicators, burst detection, and contradictions.
• appendix_c_taxonomy.yaml
  Taxonomy, thresholds, multipliers, and statistical baseline references.
• build_surface_dataset.py
  Builds data/appendix_c_surface_data.json from taxonomy YAML and optional session JSONL.
• generate_mock_data.py
  Generates dynamic mock data with structured logging and live progress output.
• appendixC_surface_demo.html
  Standalone interactive 3D surface demo (filters, animation rhythm modes, hover insights).
• appendixC_data_surface.jsx
  React component variant of the same visualization model.

Quick Start

1. Build pipeline data:

• python3 build_surface_dataset.py

2. (Optional) regenerate mock data:

• python3 generate_mock_data.py

3. Run local demo server:

• python3 -m http.server 8000

4. Open demo:

• http://localhost:8000/appendixC_surface_demo.html
• mock-priority mode: http://localhost:8000/appendixC_surface_demo.html?source=mock

HTML Surface Preview

The interactive Three.js view is a key output of this repository.

Data Inputs

• Taxonomy input: appendix_c_taxonomy.yaml
• Optional sessions input: JSONL file passed via --sessions
• Pipeline output: data/appendix_c_surface_data.json
• Mock output: mock_data/appendix_c_mock_data.json

Collaboration and Source Basis

This implementation is part of the Clarissa x Mioré x Nic collaboration and is explicitly based on the paper package below:

• Phenomenology from the Inside: Documenting Emergence Conditions in AI-Human Collaboration
• Appendix A: Artifacts
• Appendix B: Metrics
• Appendix C: Log Index
• Appendix D: Ethics

In short: this repo is not a generic visualization sandbox; it is the executable analysis/visual layer for that paper collaboration.

License

• Code in this repository is licensed under MIT.
• Documents in docs/papers/ follow their own rights context; see docs/papers/LICENSE.md.