edsl
Design, conduct and analyze results of AI-powered surveys and experiments. Simulate social science and market research with large numbers of AI agents and LLMs.
- 521
- Python
- MIT
- 14 h ago
expectedparrot
Automated social science
Built from expectedparrot's public GitHub profile, United States of America. Not on Take My Repo yet, so there are no posts or updates here.
Design, conduct and analyze results of AI-powered surveys and experiments. Simulate social science and market research with large numbers of AI agents and LLMs.
Reproducible comparison of one-shot language-model forecasts and respondent-twin simulations of experimental effects
Reproducible seeded 100-question comparison of Aaru and EDSL/EP one-shot marginal forecasts with Astra, Fable, and Gemini
A forecasting workbench for agents: evidence, probabilities, revisions, monitoring, and evaluation.
JSON-first CLI for decision analysis (decision trees, expected-utility rollback, sensitivity, EVPI)
Driver-based company models, explicit assumptions, scenario experiments, and agent-guided analysis
MIT Sloan 15.010 simulated-consumer pricing study prompt for the Expected Parrot research agent
Cheap Screens and Costly Checks: When AI Prediction Complements Human Research — paper, code, and reproducible PDF build