flesch

Formula to detect the grade level of text according to the Flesch reading ease.

Contents

• What is this?
• When should I use this?
• Install
• Use
• API
  • flesch(counts)
• Types
• Compatibility
• Related
• Contribute
• Security
• License

What is this?

This package exposes an algorithm to detect ease of reading of English texts.

When should I use this?

You’re probably dealing with natural language, and know you need this, if you’re here!

This algorithm is based on syllables, whereas some others are not, which means it’s tougher to get right and slower to calculate.

See syllable for detecting syllables.

Install

This package is ESM only. In Node.js (version 14.14+, 16.0+), install with npm:

npm install flesch

In Deno with esm.sh:

import {flesch} from 'https://esm.sh/flesch@2'

In browsers with esm.sh:

<script type="module">
  import {flesch} from 'https://esm.sh/flesch@2?bundle'
</script>

Use

import {flesch} from 'flesch'

// For “The cat sat on the mat” (1 sentence, 6 words, 6 syllables).
flesch({sentence: 1, word: 6, syllable: 6}) // => 116.14500…

// For “The Australian platypus is seemingly a hybrid of mammal and reptilian
// creature.” (1 sentence, 12 words, 23 syllables).
flesch({sentence: 1, word: 12, syllable: 23}) // => 32.50499…

API

This package exports the identifier flesch. There is no default export.

flesch(counts)

Given an object containing the number of words (word), the number of sentences (sentence), and the number of syllables (syllable) in a document, returns the reading ease associated with the document.

counts

Counts from input document.

counts.sentence

Number of sentences (number, required).

counts.word

Number of words (number, required).

counts.syllable

Number of syllables (number, required).

Returns

Result is 120 (every sentence consisting of only two one-syllable words) or lower (including negative values).

The values have the following semantics:

| Score | Semantics | | :----------: | :-------------------------------------------------- | | 90.0 – 100.0 | Easily understood by an average 11-year-old student | | 60.0 – 70.0 | Easily understood by 13- to 15-year-old students | | 0.0 – 30.0 | Best understood by university graduates |

Therefore we can use the following formula to approximate the average age a student would understand a document at, given score score:

const age = 20 - Math.floor(score / 10)

Types

This package is fully typed with TypeScript. It exports the additional type Counts.

Compatibility

This package is at least compatible with all maintained versions of Node.js. As of now, that is Node.js 14.14+ and 16.0+. It also works in Deno and modern browsers.

Related

• automated-readability — uses character count instead of error-prone syllable parser
• coleman-liau — uses letter count instead of an error-prone syllable parser
• dale-chall-formula — uses a dictionary, suited for higher reading levels
• flesch-kincaid — like flesch, returns U.S. grade levels
• gunning-fog — uses syllable count, needs POS-tagging and NER
• smog-formula — like gunning-fog-index, without needing advanced NLP
• spache-formula — uses a dictionary, suited for lower reading levels

Contribute

Yes please! See How to Contribute to Open Source.

Security

This package is safe.

License

MIT © Titus Wormer