limingrui679-design/high-stakes-analytics-decision-lab

A platform-neutral analytical Skill that profiles messy data, selects case-adaptive methods, and produces source-backed visual reports for high-stakes decisions.

What it solves

High-stakes analysis often fails due to underspecified questions, hidden data cleaning choices, and the tendency to force recommendations even when evidence is weak. This project provides an evidence-gated orchestration system that ensures rigorous data validation and reproducible evidence are established before any decision-making or action is proposed.

How it works

The system uses a strict sequence of gates and adaptive routes to move from a question to a result:

  1. Data Readiness Gate: Before analysis, the system preserves the original source, profiles quality and privacy, and checks the data contract. It can stop the workflow if critical failures (like schema or leakage issues) are found.
  2. Adaptive Routing: Depending on the question, the system routes the analysis through one of four paths: Descriptive (what is happening), Diagnostic (why it is happening), Predictive (what is likely next), or Prescriptive (what should be done).
  3. Evidence-First Output: It produces a primary "Evidence Intelligence Report" detailing methods, validation, and uncertainty. A "Decision Intelligence Brief" is only added if the evidence and context specifically justify a decision.
  4. Traceable Communication: All claims and figures are linked to machine-readable results (JSON/CSV), hashes, and rerunnable code to ensure reproducibility.

Who it’s for

Analysts and researchers performing consequential analysis where reproducibility, data integrity, and the avoidance of unfounded claims are critical.

Highlights

  • Evidence-Gated Workflow: Prevents the "upgrade" of weak evidence into strong claims by requiring data readiness and justification for each analytical route.
  • Four Analytical Routes: Structured paths for descriptive, diagnostic, predictive, and prescriptive analysis with specific discipline requirements for each.
  • Reproducible Portfolio: Includes 15 real-data cases demonstrating various evidence paths, from operational demand to spatial equity.
  • Agent Skill Integration: Can be installed as a compact "Agent Skill" for use within AI agent frameworks.

Related

  • Project
    zhnnky329/MathModeling-skills

    A rigorous workflow assistant for math-modeling contests that uses a gated pipeline and auditing layers to ensure data consistency and reproducibility while keeping modeling decisions in human hands.

  • Project
    WILLOSCAR/research-units-pipeline-skills

    A framework for agent-assisted research that transforms research goals into verifiable, resumable runs with a full evidence trail and pinned pipeline contracts.

  • Project
    hoolulu/deep-research

    An automated multi-agent system that generates institutional-grade, citable research reports in 19 languages using a four-stage search and verification pipeline.

  • Project
    danielmiessler/Substrate

    An open-source framework for documenting problems and solutions linked to authoritative data, designed to accelerate human progress through shared, evidence-based knowledge.

  • Project
    benchen4395/alpha_agentic_search

    A stable, low-latency multi-agent search framework featuring a 5-layer hierarchical memory RAG stack that allows the system to improve its performance as it is used.