Experimental Analysis Lab
Analyze
Upload a dataset, identify its columns, choose a question, and receive a structured analysis. The workflow below shows what the lab will do at each step.
01
Upload dataset
CSV to begin with, XLSX to follow. Nothing leaves the browser until you ask it to.
02
Identify columns
Types, dates, categories and identifiers detected, then confirmed by you.
03
Choose a question
Explore automatically, analyse a time series, compare groups, find relationships, forecast, or assess quality.
04
Generate analysis
Summary statistics, distributions, correlations, outliers, trend and seasonality indicators, with explanatory text.
What it will do
Row and column counts, variable types, missing values, summary statistics, distributions, correlations, outlier flags, trend detection, seasonality indicators where applicable, simple uncertainty estimates, and explanatory text alongside each result.
What it will not claim
It will not describe itself as running the historical ABCD engine, and it will not report Gaussian-process model discovery unless that is what it performed. Where an analysis is descriptive rather than inferential, it will say so.
Status
The lab is in development. It does not run the historical ABCD Automatic Statistician engine, and it does not perform Gaussian-process model discovery. When those capabilities exist here, this page will say so plainly.