From your first factor to the final report
Structure the study
Create projects and experiments with objectives, factors, units, ranges, responses, and controlled conditions. Generate a randomized factorial or Latin hypercube design.
Preserve the observations
Import CSV files, inspect types and missing values, and preserve the original text alongside a versioned dataset and source hash.
Evaluate the evidence
Run Python statistics, compare regression candidates against a mean baseline, and inspect test errors and residuals before using model estimates.
Record the next decision
Propose bounded follow-up experiments, save an experiment version, and export results with their configuration and reproducibility metadata.
The hosted workspace combines authenticated server persistence with isolated browser Python. The separate FastAPI/PostgreSQL service is available in the source for deployment; see the architecture notes.