Capabilities & Roadmap
This page is an honest statement of what Statios does today, what it does with known limitations, and what is being built next. Download the capability sheet as PDF or printable HTML.
At a glance
- 600 statistical procedures across 11 domains.
- 569 fully supported · 31 available with a documented limitation.
- Open Python engine, runs on the desktop (Tauri) and in the browser (WebAssembly/Pyodide).
- Built-in AI assistant — no API keys, no model picker; your raw data rows never leave your device (the assistant sees only schema and computed results).
- One-click export to APA tables and HTML / Markdown / LaTeX reports with citations.
What we do — fully supported
Browse the complete, citable list in the Statistical Methods Reference.
Available, with known limitations
These work today, but with a caveat worth knowing before you rely on them.
| Capability | Area | Limitation |
|---|---|---|
| Regression Discontinuity (Sharp RDD) | Regression & Generalized Linear Models | Bandwidth selection is ad-hoc (1.5*SD rule), not data-driven MSE-optimal. |
| Extended Regression Model — Probit (ERM-Probit) | Regression & Generalized Linear Models | IMR lambda SE not corrected for generated-regressor problem (Murphy-Topel correction not implemented). |
| Extended Regression Model — Ordered (ERM-Ordered) | Regression & Generalized Linear Models | Lambda SE uncorrected; no p-values reported due to generated-regressor issue. |
| Treatment effect contrasts (ATE/ATT/ATC) | Post-Estimation & Diagnostics | Uses OLS-based potential outcomes, not designed for instrumental variables. |
| Diffusion-index direct forecast (user-supplied factors) | Time Series & Macroeconometrics | NOT a true Stock-Watson Dynamic Factor Model: there is no latent-factor extraction (no PCA / Kalman / state-space estimation of common factors inside the method) — the factors must be supplied as columns. Automatic latent-factor extraction is a separate upcoming feature. |
| Unrestricted MIDAS regression | Time Series & Macroeconometrics | Currently uses same dataset for both frequencies (pending multi-dataset resolution). |
| Almon polynomial MIDAS | Time Series & Macroeconometrics | Uses single dataset for both frequencies (pending multi-dataset support). |
| Beta polynomial MIDAS | Time Series & Macroeconometrics | Single-dataset limitation; multi-dataset pending (Phase 18). |
| Exponential Almon MIDAS | Time Series & Macroeconometrics | Single dataset; multi-dataset architecture pending. |
| MIDAS nowcasting | Time Series & Macroeconometrics | Single-dataset implementation; true multi-frequency data awaits Phase 18. |
| MIDAS model comparison | Time Series & Macroeconometrics | Single-dataset; intended for separate high/low datasets pending multi-dataset manager. |
| Real Business Cycle model | Time Series & Macroeconometrics | Log-linearized first-order approximation only; no occasionally binding constraints. |
| New Keynesian 3-equation model | Time Series & Macroeconometrics | Simplified linearized form; no ELB, expectations-driven dynamics approximate. |
| DSGE calibration | Time Series & Macroeconometrics | Moment-matching heuristic only; not Bayesian estimation or SMM. |
| Regression Discontinuity (Sharp RDD) | Panel Data & Causal Inference | Bandwidth selection is ad-hoc (not Imbens-Kalyanaraman or Calonico-Cattaneo-Titiunik data-driven optimal); local-linear only. |
| Propensity Score Matching (PSM) | Panel Data & Causal Inference | Uses naive paired SE; does not apply Abadie-Imbens (2006) variance correction for estimated scores and matching with replacement. |
| Hierarchical Agglomerative Clustering | Multivariate, Factor & Latent-Variable Analysis | O(n²) memory constraint; refuses n > 20,000 rows. |
| IRT Fit Indices (Approximate) | Multivariate, Factor & Latent-Variable Analysis | Approximate method: fixed-1PL reference and diagonal weight matrix, not full Maydeu-Olivares M2. |
| Shared Frailty Cox Model | Survival, Bayesian, Meta-Analysis & Resampling | Gamma frailty only; variance estimated via martingale residuals. |
| Bayesian Hierarchical Model | Survival, Bayesian, Meta-Analysis & Resampling | Fallback uses OLS-shrinkage approximation without full MCMC; ICC computed from variance components. |
| Bayes Factor | Survival, Bayesian, Meta-Analysis & Resampling | BIC-based approximation; not exact marginal likelihood. |
| SHAP Values | Machine Learning & Choice Models | TreeExplainer unavailable in browser (WASM); fallback to KernelExplainer (sampled Shapley values, not exact TreeSHAP). |
| CHAID | Machine Learning & Choice Models | Stub implementation; returns placeholder. |
| QUEST | Machine Learning & Choice Models | Stub implementation; returns placeholder. |
| C5.0 | Machine Learning & Choice Models | Stub implementation; returns placeholder. |
| Ensemble Trees Comparison | Machine Learning & Choice Models | Stub implementation; returns placeholder. |
| Tree Export | Machine Learning & Choice Models | Stub implementation; returns placeholder. |
| Compare Trees | Machine Learning & Choice Models | Stub implementation; returns placeholder. |
| Hierarchical Bayes CBC | Machine Learning & Choice Models | Empirical-Bayes / EM-style scheme, NOT Gibbs/Metropolis; samples are posterior modes, not draws; heterogeneity SD understates posterior uncertainty. |
| Session Undo | Data Management, Transforms & Workflow | Session-level undo not fully wired; edit.undo handles cell-level reversals. |
| Session Redo | Data Management, Transforms & Workflow | Session-level redo not fully wired; edit.redo handles cell-level reversals. |
| Large-dataset/columnar mode: bounded analysis suite | data | WASM in-browser path capped at 1.5M rows or 900MB. CSV files up to 50M rows route to columnar store (Parquet/NPY), but only subset of analyses support column-on-demand (streaming). Binary formats (.xlsx/.sav/.dta) cannot be streamed. |
| WASM/Pyodide compatibility: numba and SHAP slower without JIT | ml | Web build stubs numba (Pyodide lacks LLVM JIT) and SHAP's _kernel_lib. Code runs interpreted (correct numerically, ~2–5× slower than compiled). KernelExplainer works; TreeExplainer unavailable in WASM. |
| Charts: 4 types shipped, 31 roadmap items incomplete | graphs | Implemented: histogram, scatter, boxplot, bar. Missing: violin, density, Q–Q, interaction, forest, ROC, funnel, survival curves, residual diagnostic panels, and 21 others from Phase 15 roadmap. |
| Extended regression diagnostics: some estimators in pre-release rigor improvements | regression | Arellano-Bond, GMM, DSGE solvers present but some advanced econometric estimators have rigor enhancements on unmerged branches. Shipped versions work; see code docstrings for limitations (e.g., 'simplified solver', 'bounded lag range'). |
In active development
| Item | Area | Notes |
|---|---|---|
| Statios web pivot: Python engine to Pyodide/WASM | infrastructure | W1–W4 phases (June 2026): migrate sidecar to in-browser Pyodide via WorkerTransport. Desktop Tauri shell frozen. PWA deployment, OPFS file handling, web deploy pipeline. Completed handler count remains 445; no new features, transport refactor only. |
| Charts Expansion: 31 missing chart types | graphs | Violin, density (KDE), ridgeline, Q–Q, P–P, interaction plot, mosaic, ACF/PACF, IRF with CI, residual diagnostics (4-panel), forest/funnel, survival, ROC, Bland-Altman, coefficient, marginal effects, control chart, caterpillar. Phase 15 completion. |
| AI Orchestrator & Multi-Model Routing | ai | Post-W4: intelligent provider selection, prompt caching, fallback logic, cost tracking. Statios relay authentication for managed web tier. |
| Quality backlog: UI refinements, performance, test coverage | polish | Post-W4: Monaco editor for scripts, keyboard shortcuts, undo/redo on data transforms, canvas DataGrid for >10k rows, performance audits on large-dataset analyses. |
On the roadmap
Longer-horizon work — much of it has a working engine implementation already and is awaiting UI surfacing, polish, or expanded test coverage.
| Item | Area | Notes |
|---|---|---|
| Advanced causal inference: bounds & sensitivity analyses | causal | Lee bounds, Rosenbaum sensitivity, E-values. Handlers present (causal.lee_bounds, etc.) but backlog for UI polish and help content. |
| Exotic time-series models: TAR/STAR, NARDL, nowcasting | timeseries | Handlers exist (ts.tar, ts.nardl, ts.nowcasting) but limited test coverage and optional dependencies (ruptures for Bai–Perron). Deferred to post-quality phase. |
| Multivariate: MANOVA, discriminant, canonical correlation | analysis | Handlers registered (test.manova, analysis.discriminant, analysis.canonical_correlation) but frontend UI tabs not yet exposed in AnalysisModal. Feature-complete backend; UI routing needed. |
| Bayesian posterior predictive checks & Bayes factors | bayes | Handlers: bayes.ppc, bayes.credible_interval, bayes.prior_posterior, bayes.bayes_factor. Deps (PyMC, ArviZ) optional. Lower priority; posterior work focuses on MCMC diagnostics. |
| Optimal scaling & ERM: categorical regression (CATREG), optimal transforms | analysis | Handlers: analysis.catreg, analysis.catpca, analysis.overals, analysis.catcorr. Phase 23 backends exist but semi-automated UI guidance and diagnostic plots deferred. |