Overview

A bunch of functions to analyse frequencies in text.

Installation

npm install text-frequencies-analysis

Tuples

This collection of functions is all about tuples. Here, a tuple means an array with two items. The first item is a string, and the second one a number—the frequency of the string. For example, ["a",5].

There are two functions—countEach and countPairs—that return arrays of tuples. In other words, they produce data.

The rest of the functions let you work with the data.

Usage

Each function is described in the source code. The tests are also informative.

Moreover, the file analyse.js is an example on how to use the tuple-producing functions. Here’s a gist of the output, as an example.

Here’s a small taste of what you can do with the rest of the functions:

filter lets you filter an array of tuples. For example, to filter out all pairs of english letters that contain the letter “e” and another vowel: from an array of character-tuples:

filter(pairs, /e/, /[aouiy]/)

relative makes frequencies into percentages.

sumTuples sums all the frequencies in an array of tuples.

If we wanted to know how many percent the “e+vowel” pairs make out of all pairs, we could run the following:

sumTuples(filter(relative(pairs), /e/, /[aouiy]/))

Sometimes it’s hard to believe that a certain pair of letters is as common as your data say. Then it’s useful to get a list of words that include that pair.

pairWords(words, "ul")

License

All files are in the public domain.