pymc-labs

pymc-labs / decision-lab

Public

Run tested, autonomous agent workflows on your data for meaningful decision-making

27
2
100% credibility
Found Apr 01, 2026 at 27 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

decision-lab runs AI agents with domain skills and parallel subagents in locked environments to produce robust data analysis reports by checking convergence across approaches.

How It Works

1
🔍 Discover decision-lab

You hear about a smart tool that helps AI assistants do reliable data analysis by trying many approaches at once.

2
📦 Pick a ready-made analysis kit

Choose a pre-made kit for your type of work, like marketing or forecasting, that has all the right instructions and tools.

3
📊 Add your numbers and question

Share your data spreadsheet and ask your question, like 'What's working best?'

4
🚀 Start the analysis

Hit go, and it sets everything up safely and runs smart helpers in parallel to check different ideas.

5
📱 Watch it think live

Peek in real-time to see the helpers working together and tracking costs.

6
Review the combined wisdom
✅
Results match

Get solid recommendations you can trust.

❓
Mixed results

Learn uncertainties and next steps.

🎉 Confident decisions

Use the clear report with charts and advice to make better choices.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 27 to 27 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is decision-lab?

decision-lab runs autonomous AI agent workflows on your datasets to deliver scrutinized decisions, like budget optimizations from marketing data. You bundle prompts, skills, and pinned Docker environments into portable "decision-packs," feed in CSV data via CLI (`dlab --dpack mmm --data sales.csv --prompt "recommend allocations"`), and get reports, plots, plus convergence checks across parallel modeling paths. Built in Python with Modal cloud support, it catches unreliable inferences early, unlike solo agents that commit to one approach.

Why is it gaining traction?

It solves AI agents' "forking paths" problem in data science—single models often mislead, but decision-lab deploys parallel subagents (diverse LLMs/strategies), consolidates via tools, and flags divergence, yielding trustworthy outputs. User perks include live TUI monitoring (`dlab connect`), Gantt timelines (`dlab timeline`), creation wizards, and Decision Hub skills integration. Reproducible Docker runs and GitHub workflow-like execution (run locally or on branches) make agentic analysis production-ready.

Who should use this?

Data scientists automating Bayesian MMM or forecasting who distrust vanilla agent code. Marketers needing robust ROAS insights without manual convergence checks. Teams at places like decision lab uni köln or esade decision lab building decision lab promethee-style multi-criteria tools, or devs wanting to run github actions locally for agent testing before cloud deploys.

Verdict

Promising alpha from PyMC Labs (27 stars, 1.0% credibility score) with strong docs and MMM examples, but low traction means watch for stability. Prototype a decision-pack now if agent workflows frustrate you—CLI delivers fast wins despite early maturity.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.