tdigest

What is tdigest?

The tdigest npm package is a JavaScript implementation of Ted Dunning's t-digest algorithm, which is used for accurate online accumulation of rank-based statistics such as quantiles and cumulative distribution functions (CDFs). It is particularly useful for large-scale data processing where memory efficiency and speed are critical.

What are tdigest's main functionalities?

Creating a t-digest

This feature allows you to create a new t-digest instance, which can then be used to accumulate data and compute statistics.

const TDigest = require('tdigest').TDigest;
const t = new TDigest();

Adding data to the t-digest

This feature allows you to add data points to the t-digest. The t-digest will then use these data points to compute rank-based statistics.

t.push(1.0);
t.push(2.0);
t.push(3.0);

Computing quantiles

This feature allows you to compute quantiles from the data accumulated in the t-digest. For example, you can compute the median (50th percentile) or the 90th percentile.

const q50 = t.percentile(0.5);
const q90 = t.percentile(0.9);

Computing cumulative distribution function (CDF)

This feature allows you to compute the cumulative distribution function (CDF) for a given value. The CDF represents the probability that a random variable takes on a value less than or equal to the given value.

const cdf = t.cdf(2.0);

Other packages similar to tdigest

simple-statistics

The simple-statistics package provides a wide range of statistical functions, including quantile calculations. However, it does not implement the t-digest algorithm and may not be as memory efficient for large-scale data processing.

summary-statistics

The summary-statistics package offers basic statistical summaries such as mean, median, and standard deviation. It does not provide the advanced rank-based statistics or memory efficiency of the t-digest algorithm.

descriptive-statistics

The descriptive-statistics package provides functions for descriptive statistics, including quantiles. While it offers similar functionality, it does not use the t-digest algorithm and may not be as efficient for large datasets.

README

tdigest

tdigest: javascript implementation of Dunning's T-Digest for streaming quantile approximation