Co-created causal knowledge
An editorial framework for turning knowledge from people and communities into documented questions, triage decisions, and interpretation boundaries.
Public learning library
Learn the architecture without entering the project workspace.
CO-LUMINATE connects co-created causal stories, longitudinal reconstruction, outcome harmonisation, causal specification, and governed evidence release. These project-owner-directed editorial summaries contain no participant-level data and are not released artifacts.
Platform horizons
Ambitious enough to grow, explicit enough not to overclaim.
CO-LUMINATE is intended to support multiple populations, outcomes, and implementation settings over time. Maturity badges distinguish what can be learned from now from work still being designed.
Learning and transparent governance
Project-owner-directed editorial summaries of synthetic-data rules, reproducibility guidance, and public/private handover safeguards.
Multi-population and multi-outcome adaptation
A governed adaptation workflow is being defined. No population or outcome beyond a human-approved implementation should be considered solved.
Reusable synthetic-first pipelines
Internal specifications and synthetic reference files exist. No licensed, runnable, or reusable pipeline is published on this site.
Evidence to decision and implementation
The roadmap connects bounded evidence to learning, implementation partnerships, and action while keeping uncertainty visible.
Capability-scoped extensions and plugins
Future integrations may add narrowly permissioned tools and services. No plugin marketplace or runnable extension is available today.
Start-to-finish learning path
Follow the decisions in the order they become defensible.
Each stage should leave an auditable basis for the next. Later modelling cannot repair undocumented upstream choices.
- 01Frame
Begin with lived experience, purpose, and an explicit causal question.
- 02Reconstruct
Make linkage, time, households, exposures, and observation gaps visible.
- 03Harmonise
Document endpoint provenance, measurement limits, and adaptation needs.
- 04Specify
Translate an approved DAG into requirements, assumptions, and an estimand.
- 05Validate
Test contracts, provenance, failure modes, and interpretation boundaries.
- 06Release
Share only reviewed, licensed, public-safe learning artifacts.
Current public summaries
Learn by question, not by repository folder.
Each card is a current public editorial summary. It is not a download, runnable tool, pipeline, or substitute for the internal source material.
01
Current editorial summaryArchitecture
Platform foundations
Understand why CO-LUMINATE is a reusable evidence architecture, with depression as a reference case rather than the platform boundary.
What this summary covers
The five layers, module boundaries, and separation between public and controlled work.
02
Current editorial summaryCo-creation
From causal stories to measurable questions
Learn how lived-experience inputs can inform causal questions without being treated as automatically model-ready variables.
What this summary covers
How causal stories can be captured, triaged, and interpreted without turning them directly into model-ready variables.
03
Current editorial summaryLife-course design
Life-course reconstruction
Work through person-time, visits, households, exposure windows, linkage quality, and observation gaps before analysis.
What this summary covers
The sequence of timing, linkage, history, and gap questions that should be resolved before causal modelling.
04
Current editorial summaryOutcome adaptation
Outcome harmonisation and adaptation
See which provenance, timing, measurement, missingness, and validation decisions must be revisited for every new outcome.
What this summary covers
The decisions that must be revisited before adapting the architecture beyond the reference outcome.
05
Current editorial summaryCausal reasoning
Causal reasoning workflow
Move from a documented question and DAG to analysis windows, estimands, interpretation limits, and sensitivity needs.
What this summary covers
The assumptions, timing decisions, estimand questions, and interpretation limits that precede analysis.
06
Current editorial summaryGovernance and reproducibility
Reproducibility and governed handover
Learn how provenance, environments, tests, validation reports, synthetic-data rules, and release review fit together.
What this summary covers
The provenance, testing, privacy, and human-review controls required before a public handover.
07
Current editorial summaryResearch and roadmap
Research evidence and implementation roadmap
Review the evidence-to-decision basis, architectural recommendations, risks, and staged implementation plan.
What this summary covers
The cited research synthesis, recorded risks, and staged implementation roadmap.
08
Current editorial summaryAgent governance
Responsible AI-assisted development
See how bounded tasks, source-of-truth rules, privacy checks, and explicit open questions reduce unsafe automation.
What this summary covers
The source-of-truth, privacy, review, and open-question controls applied to AI-assisted development.
Reuse readiness
Read the public summaries now. Source artifacts are not published for redistribution.
- License
- Pending project-owner decision
- Release approval
- Pending real release manifest and human approval
- Downloads
- No public downloads
Public editorial learning summaries are available on this site. No tool, pipeline, source package, download, or redistribution permission is currently published. Publication requires an approved license, completed validation and safety review, a real release manifest, and human approval.