How to compute multiple statistics at once in kixi.stats?

Use the `fuse` combinator from the redux library to combine reducing functions, as shown in the README for calculating mean and standard deviation simultaneously with `transduce`.

kixi.stats vs Incanter for Clojure statistics?

kixi.stats focuses on transducer-based, composable statistics and sampling, while Incanter offers a broader suite including charts and more advanced models. Choose kixi.stats for pipeline integration and Incanter for richer statistical environments.

How to generate repeatable random samples in kixi.stats?

Pass a seed value in an options map to functions like `draw` or `sample`, e.g., `(draw distribution {:seed 42})` ensures deterministic results, useful for reproducible simulations.

Does kixi.stats work with ClojureScript for web apps?

Yes, it's cross-platform, but note that some features like t-digest histograms are optimized for Clojure; check the documentation for any ClojureScript-specific limitations.

How to calculate quantiles from streaming data in kixi.stats?

Use the `histogram` function with `post-complete` to build an empirical distribution and query quantiles efficiently, as demonstrated in the README for avoiding multiple digest computations.

What hypothesis tests are available in kixi.stats?

It includes Z-tests (one-sample and two-sample), t-tests (simple and Welch's), and chi-squared tests for categorical independence, accessible via the `kixi.stats.test` namespace or as reducing functions.

kixistats

Clojurev0.5.8

A Clojure/ClojureScript library for statistical distribution sampling and transducing functions.

Visit Website

What is kixistats?

kixi.stats is a statistical library for Clojure and ClojureScript that provides functions for sampling from probability distributions and computing statistics via transducers. It enables efficient, composable statistical computations that integrate seamlessly with Clojure's core data processing workflows. The library solves the problem of performing statistical analysis in a way that aligns with Clojure's functional and data-centric paradigms.

Target Audience

Clojure and ClojureScript developers who need to perform statistical analysis, sampling, or hypothesis testing within data processing pipelines. This includes data scientists, engineers building analytical applications, and developers working on simulations or data-intensive systems in the Clojure ecosystem.

Value Proposition

Developers choose kixi.stats because it deeply integrates statistical functions into Clojure's transducer ecosystem, allowing statistical computations to be composed flexibly with existing data transformations. Its unique selling point is providing a comprehensive statistical toolkit—including distribution sampling, hypothesis tests, and empirical analysis—that works natively with Clojure's reducing functions and transducers, avoiding monolithic APIs.

Overview

A library of statistical distribution sampling and transducing functions

Use Cases

Best For

Related Projects

Community-curated · Updated weekly · 100% open source

Found a gem we're missing?

Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.

Submit a project Star on GitHub

GitHub

368 stars20 forks0 contributors

Sampling from over 20 probability distributions (e.g., Normal, Binomial, Poisson, Gamma) in Clojure/ClojureScript applications.
Computing statistics like mean, variance, correlation, and regression using transducers within data processing pipelines.
Performing hypothesis testing with Z-tests, t-tests (including Welch's), and chi-squared tests in Clojure.
Creating and querying empirical distribution histograms using t-digest for quantiles, median, and interquartile range analysis.
Generating repeatable, deterministic samples by providing seed values for reproducible simulations or experiments.
Efficiently summarizing discrete distributions (e.g., Bernoulli, categorical) without intermediate sample reification for performance-critical applications.

Not Ideal For

Projects built in languages other than Clojure or ClojureScript that require statistical libraries.
Teams needing interactive, GUI-based statistical software for exploratory data analysis or visualization.
Applications requiring advanced statistical models like multivariate regression, ANOVA, time series analysis, or Bayesian inference.

Pros & Cons

Pros

Seamless Transducer Integration

Statistical functions like mean and standard deviation are reducing functions compatible with `transduce`, allowing composable computations within Clojure's data pipelines, as shown in examples combining statistics with filters.

Comprehensive Distribution Sampling

Supports over 20 probability distributions including Normal, Binomial, and Poisson, enabling diverse simulation and sampling needs directly in code.

Efficient Empirical Analysis

Uses t-digest for histograms to compute quantiles, median, and interquartile range efficiently, with optimizations like `histogram` and `post-complete` to avoid redundant calculations.

Deterministic and Reproducible

Allows repeatable sampling by providing seed values in functions like `draw` and `sample`, crucial for testing and simulations, as demonstrated in the deterministic sampling section.

Cons

Limited Advanced Statistical Scope

Focuses on basic distributions, descriptive stats, and hypothesis tests; lacks support for more complex methods like multivariate analysis, machine learning, or Bayesian statistics, which might require additional libraries.

Dependency on External Combinators

For composing multiple statistics in one pass, it relies on third-party libraries like redux or xforms (e.g., using `fuse`), adding dependencies and complexity beyond core Clojure.

Clojure-Specific Ecosystem Constraints

Tied to Clojure/ClojureScript, it has a smaller community and fewer integrations compared to mainstream statistical tools in Python or R, limiting collaboration and resource availability.

Frequently Asked Questions

Home

Machine Learning

promptfoo

Test your prompts, agents, and RAGs. Red teaming/pentesting/vulnerability scanning for AI. Compare performance of GPT, Claude, Gemini, DeepSeek, and more. Simple declarative configs with command line and CI/CD integration. Used by OpenAI and Anthropic.

💡 All-in-one AI framework for semantic search, LLM orchestration and language model workflows

Kedro is a toolbox for production-ready data science. It uses software engineering best practices to help you create data engineering and data science pipelines that are reproducible, maintainable, and modular.

Production ready toolkit to run AI locally

#probability-distributions

Machine Learning72.2k