The Stats Duck
A statistical computing toolkit for DuckDB.
The Stats Duck adds statistical analysis to SQL: descriptive statistics,
hypothesis tests, grammar-of-graphics visualization, and direct readers/writers
for SAS, SPSS, and Stata files. Functions are implemented as streaming
aggregates and scalar primitives, so they scale from local notebooks to
billion-row warehouses and also run inside DuckDB-WASM in the browser.
The Stats Duck is built by KoliStat (see the
product page). It is the
statistical engine of Bedevere Wise, KoliStat's
browser-based statistical workbench.
The extension installs and loads in DuckDB under the technical name
stats_duck (matching the binary, the SQL function namespace, and the
INSTALL keyword). "The Stats Duck" is the project / brand name; stats_duck
is what you type at the SQL prompt.
The name is a play on The Fat Duck, Heston
Blumenthal's restaurant in Bray, Berkshire. That restaurant has been the
inspiration for the project name from the start.
Scope
The Stats Duck is meant to cover the everyday work of a general-purpose
statistician without leaving SQL. The current release covers four areas:
- Hypothesis tests — parametric and non-parametric, with effect sizes and
confidence intervals returned alongside the test statistic.
- Visualizations (VISUALIZE) —
VISUALIZE … FROM <table> DRAW <mark>, a
Posit-published Grammar-of-Graphics SQL dialect compiled to Vega-Lite v5.
No server-side rendering: the extension emits a spec + per-layer SQL, and
the client (browser, notebook, …) runs the SQL and feeds the rows to vega.
- Statistical file I/O — readers and writers for SAS, SPSS, and Stata
files, integrated with DuckDB's virtual file system so they work with local
paths,
httpfs://, s3://, and registered WASM file buffers.
- Streaming aggregates — every test is a single-pass aggregate over the
data, so it works with
GROUP BY, window frames, and DuckDB's
parallel execution.
Future releases will add Spearman/Kendall correlations, regression with full
diagnostics, multiple-testing corrections, and more distribution families.
Further documentation:
Functions
Hypothesis tests (aggregate)
| Function | Description |
|---|
ttest_1samp(column, [mu], [alpha], [alternative]) | One-sample t-test |
ttest_2samp(column1, column2, [equal_var], [alpha], [alternative]) | Two-sample t-test (Welch's or Student's) |
ttest_paired(column1, column2, [alpha], [alternative]) | Paired t-test |
mann_whitney_u(column1, column2, [alternative], [continuity]) | Mann-Whitney U test (Wilcoxon rank-sum) |
wilcoxon_signed_rank(column1, column2, [alternative], [continuity]) | Wilcoxon signed-rank test |
pearson_test(x, y, [alpha], [alternative]) | Pearson correlation with significance |
spearman_test(x, y, [alpha], [alternative]) | Spearman rank correlation |
kendall_test(x, y, [alternative]) | Kendall's tau-b rank correlation |
anova_oneway(value, group) | One-way ANOVA |
chisq_independence(row, col, [continuity]) | Chi-square test of independence |
chisq_goodness_of_fit(category) | Chi-square goodness-of-fit (uniform) |
jarque_bera(column) | Jarque-Bera normality test |
shapiro_wilk(column) | Shapiro-Wilk normality test (Royston AS R94) |
anderson_darling(column) | Anderson-Darling normality test |
ks_test_1samp(column) | Kolmogorov-Smirnov one-sample (vs fitted normal) |
ks_test_2samp(column1, column2) | Kolmogorov-Smirnov two-sample |
sign_test_1samp(column, [mu], [alternative]) | Sign test on the median |
sign_test_paired(column1, column2, [alternative]) | Paired sign test |
All tests return a STRUCT with the test statistic, degrees of freedom,
p-value, and relevant effect sizes / confidence intervals.
Every test skips NULL inputs, and treats NaN the same way: a NaN value
counts as missing and never enters the test. For two-column tests a NaN
on either side drops the pair (paired tests) or that single value
(two-sample tests), exactly as a NULL would.
This is the beginning of the README. Read the rest at the original. Read the full README ↗
