Real statistical analysis usually means a heavyweight desktop tool or a blank spreadsheet
Between SPSS-grade software that’s overkill for a single test and a bare spreadsheet with no guidance, there’s little in between for someone who just wants to run a real statistical test against real data quickly.
A workbench built on a real statistics engine
StatLab is a React front end powered by its own published statistics engine, `statlab`, keeping the math logic independent of the UI and testable on its own.
Per-test dataset recommendations — including bundled psychometrics datasets like a Life Satisfaction Survey and a Vocabulary Test — drive a recommendations panel so every test starts from a meaningful example.
Core surfaces that define the experience
Pick a test from a modular library
The test library is organized so choosing the right statistical test doesn’t require already knowing its name — the interface surfaces what each test is for.
Built for people running the analysis, not just teaching it.
Recommended datasets, not an empty grid
A per-test recommended-datasets mapping surfaces relevant example data, including bundled real-world datasets, so the calculator never opens to a blank state.
Every test has somewhere real to start.
Worked results from a real engine
Results come from the standalone statlab engine rather than inline formulas, so the calculation logic is shared, tested, and reusable outside the app.
The math is a package, not a UI implementation detail.
What makes it work
A Real Engine, Not a Demo
Statistical tests run against a standalone, npm-published engine package rather than a hardcoded set of formulas baked into the UI, so the math is reusable and independently testable.
Dataset Recommendations Per Test
Each test advertises which bundled datasets it works well with, so choosing a test surfaces a meaningful example instead of an empty grid waiting for input.
Fully Client-Side
All computation runs in the browser against the statlab engine — no data ever leaves the device, which matters for anyone working with sensitive research data.
How a calculation runs
StatLab is built around picking a test, loading data, and reading real results.
01
Choose a statistical test
A modular test library makes the right test discoverable without prior expertise.
02
Load a recommended or custom dataset
Per-test dataset recommendations, including bundled datasets, give every test a meaningful starting point.
03
Run it against the statlab engine
Computation happens client-side against the published statlab engine package — no upload required.