A GUIDED TOUR · CAPTURED LIVE FROM THE APP

See the workbench in motion.

Every screen below is the real Statios — running entirely in the browser, computing on your device. From raw data to a cited, reproducible, publication-ready result, without a single upload.

statios-app.web.app · engine: ready · n=140
Lobby — start screen Data view Variable view Regression result Analysis breadth Chart Guided path Reproducible code Report Statistics Wiki
▸ auto-playing · ten real screens, on loop
600
procedures
220
validated vs R
315
wiki articles
100%
runs on-device
00 · THE LOBBY

Your projects, right where you left them.

Sign in and land in the Lobby — a clean start screen showing your recent projects, account status and usage. Open an existing project or start a new one with a single click, then the full workbench loads instantly.

  • New / Open — one click to create or resume a project
  • Recent projects listed with last-modified timestamps
  • Account chip and usage summary always in reach
New
Statios · Lobby · New · Open · Recents
Lobby start screen
01 · THE DATA ENGINE

A spreadsheet that thinks in statistics.

Open a CSV, Excel, SPSS .sav or Stata .dta and it's instantly live — a fast, windowed grid with a formula bar, type-aware columns and inline editing.

  • Hundreds of thousands of rows scroll fluidly; the engine reaches into the millions
  • Excel-style formulas, transforms, recodes and computed columns
  • Full undo/redo — every edit is tracked
Data View · clinical_trial · 140×8
Data grid with formula bar
02 · VARIABLE LABELING

SPSS-style metadata, done right.

A proper Variable View: set type, measurement level, value labels and roles — the metadata real analysis depends on. Continuous fields get one-click Visual Binning.

  • Value labels & auto-encode categoricals for modelling
  • Visual Binning — cut a scale into groups with live previews
  • Measurement levels flow through every downstream test
New
Variable View · type · measure · labels · bin
Variable view with binning
03 · EVIDENCE, NOT VIBES

Every result arrives cited.

Run an analysis and you get a clean, APA-style card — coefficients, robust SEs, effect sizes, confidence intervals and significance — with a Cite button on every output.

  • One click copies the methodological references behind the result
  • The numeric engine is validated against R on a committed golden-fixture suite
  • Diagnostics and assumption checks travel with the estimate
Linear Regression (OLS) · R² · 95% CI · Cite
Cited regression output
04 · ONE WORKBENCH

The scope of five tools, in one menu.

Regression families, multivariate & latent-variable models, time series & macroeconometrics, panel data, causal inference, survival, Bayesian, machine learning — 600 procedures, all on-device.

  • SPSS + Stata + EViews coverage without four licenses
  • Every method computed by an open, auditable Python engine
  • Native .sav / .dta / Excel / CSV — read and written
Analyze · 600 procedures
The analyze menu showing breadth
05 · VISUALIZATION

37 chart types, publication-clean.

Build grouped scatter, distributions, box and violin plots, correlation heatmaps, survival curves, forest plots and more — themed to match, interactive on hover.

  • Group, facet and color by any variable
  • High-res export to PNG (300dpi) or vector SVG
  • Dark / light theme-aware palettes
Graphs · cholesterol vs age, by group
Grouped scatter plot
06 · GUIDED PATH

Method choice grounded in literature.

"Which test should I run?" — answered by a versioned knowledge base of decision rules drawn from published methodology. The AI only interviews you; the rules choose the methods, and every step ships with citations.

  • Describe your study in plain words; get a defensible plan
  • Deterministic planner — same study, same plan, reviewable
  • Export the cited justification for your methods section
New
Guided · plan your analysis
Guided path planning screen
07 · THE STATISTICS WIKI

A statistics reference, built in.

Every method comes with a plainly-written, PhD-depth explainer — what it does, when to use it, the assumptions and the math — fully cited, and one click from running it. 315 articles covering all 600+ procedures.

  • Searchable and cross-linked, with the source each method comes from
  • Open in Statios — jump from an article straight to the dialog that runs it
  • No tab-switching to Google — the reference lives beside your work
New
Wiki · every method, explained & cited
Statistics Wiki article
08 · REPRODUCIBLE BY DEFAULT

Your whole session, as a script.

Everything you click is recorded as a deterministic Python script — a faithful replay of the exact engine calls, in order. Not AI-generated: each call is rendered from a per-method template into real Python (statsmodels / scipy / lifelines / …), runnable outside Statios with only open-source deps.

  • Copy or download .py — ready for a paper's code appendix
  • Re-run it to regenerate every figure on the same data
  • Pairs with the report's APA reference list
New
Lab → Reproduce · deterministic replay · download .py
Reproducible code view
09 · REPORTING

From clicks to a finished document.

Assemble a typeset Statistical Analysis Report — a deterministic layout of data, methods, APA results, tables and figures, ending in a reproducible-code appendix with a complete reference list.

  • Export to PDF or Markdown in a click
  • Automatic References section — nothing to assemble by hand
  • "Prepared with Statios — deterministic statistics for research"
Report · Statistical Analysis Report — clinical_trial
Publication-ready report

That's the lab. Open it.

No install, no upload — a free account lets you start in seconds, and your data stays on your machine.

Launch Statios →