kstats

A Kotlin Multiplatform statistics toolkit covering descriptive analysis, probability distributions, hypothesis testing, correlation, regression, and sampling. Pure Kotlin, published to Maven Central as focused modules.

Quickstart

val data = doubleArrayOf(2.0, 4.0, 4.0, 5.0, 7.0, 9.0)
val summary = data.describe()
// => DescriptiveStatistics(count=6, mean=5.17, median=4.5, standardDeviation=2.48, ...)

data.mean()               // => 5.1667
data.standardDeviation()  // => 2.4833
data.skewness()           // => 0.3942

Installation

Gradle (BOM)

dependencies {
    implementation(platform("org.oremif:kstats-bom:0.5.0"))
    implementation("org.oremif:kstats-core")
    // add other modules as needed
}

Kotlin Multiplatform

kotlin {
    sourceSets {
        commonMain.dependencies {
            implementation(project.dependencies.platform("org.oremif:kstats-bom:0.5.0"))
            implementation("org.oremif:kstats-core")
        }
    }
}

If you only need one module: implementation("org.oremif:kstats-core:0.5.0").

Modules

Module	Description
kstats-core	Descriptive statistics, moments, quantiles, streaming stats
kstats-distributions	28 probability distributions (18 continuous + 10 discrete)
kstats-hypothesis	Parametric, non-parametric, normality, and categorical tests
kstats-correlation	Correlation coefficients, covariance matrices, simple linear regression
kstats-sampling	Ranking, normalization, binning, bootstrap, weighted sampling

kstats-core

val data = doubleArrayOf(2.0, 4.0, 4.0, 5.0, 7.0, 9.0)
val summary = data.describe()
summary.mean              // => 5.1667
summary.median            // => 4.5
summary.standardDeviation // => 2.4833

val stats = OnlineStatistics()
stats.addAll(doubleArrayOf(1.0, 2.0, 3.0, 4.0, 5.0))
stats.mean                // => 3.0
stats.standardDeviation() // => 1.5811

kstats-distributions

val normal = NormalDistribution(mu = 0.0, sigma = 1.0)
normal.pdf(0.0)                  // => 0.3989
normal.cdf(1.96)                 // => 0.9750
normal.quantile(0.975)           // => 1.9600
normal.sample(5, Random(42))     // => [0.11, -0.87, ...]

kstats-hypothesis

val sample = doubleArrayOf(2.0, 4.0, 4.0, 5.0, 7.0, 9.0)
val result = tTest(sample, mu = 5.0)
result.statistic                 // => 0.1644
result.pValue                    // => 0.8759
result.isSignificant(alpha = 0.05) // => false

kstats-correlation

val x = doubleArrayOf(1.0, 2.0, 3.0, 4.0, 5.0)
val y = doubleArrayOf(2.1, 3.9, 6.2, 7.8, 10.1)

val r = pearsonCorrelation(x, y)
r.coefficient                    // => 0.9987
r.pValue                         // => 0.0001

val reg = simpleLinearRegression(x, y)
reg.slope                        // => 1.99
reg.rSquared                     // => 0.9973
reg.predict(6.0)                 // => 11.99

kstats-sampling

val data = doubleArrayOf(3.0, 1.0, 4.0, 1.0, 5.0)
data.rank()                      // => [3.0, 1.5, 4.0, 1.5, 5.0]
data.zScore()                    // => [-0.16, -1.47, 0.49, -1.47, 1.14]

listOf(1, 2, 3, 4, 5).bootstrapSample(10, Random(42))

val dice = WeightedDice(mapOf("A" to 3.0, "B" to 1.0))
dice.roll()                      // => "A" (75% probability)

Platform Support

kstats targets every platform Kotlin supports.

Platform	Targets
JVM	jvm
Android	android, androidNativeArm32, androidNativeArm64, androidNativeX86, androidNativeX64
iOS	iosArm64, iosX64, iosSimulatorArm64
macOS	macosArm64
watchOS	watchosArm32, watchosArm64, watchosDeviceArm64, watchosSimulatorArm64
tvOS	tvosArm64, tvosSimulatorArm64
Linux	linuxArm64, linuxX64
Windows	mingwX64
JS	js
Wasm	wasmJs, wasmWasi

Documentation

• Guides and tutorials: kstats.oremif.org
• API reference: Dokka site
• Interactive notebooks: samples/ — Kotlin Notebooks with Kandy visualizations for each guide
• Benchmarks: benchmark/README.md — JMH results comparing kstats vs Apache Commons Math

Development

./gradlew jvmTest        # run JVM tests
./gradlew allTests       # run all platform tests
./gradlew build          # full build

./gradlew :benchmark:benchmark      # JMH benchmarks (kstats vs Apache Commons Math)
./gradlew :benchmark:smokeBenchmark # quick smoke run

Contributing

Contributions are welcome. See CONTRIBUTING.md for the contribution workflow, issue guidelines, and project conventions. Please also read CODE_OF_CONDUCT.md before participating in issues, pull requests, or discussions.

License

Licensed under the Apache License 2.0.