Skip to content

Add quartiles, interquartiles and outliers functions #7

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
vanakenm wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Add quartiles, interquartiles and outliers functions #7

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 17 additions & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -107,6 +107,23 @@ Calculate the [standard deviation](http://en.wikipedia.org/wiki/Standard_deviati

Calculate the value representing the desired [percentile](http://en.wikipedia.org/wiki/Percentile) `(0 < ptile <= 1)`. Uses the Estimation method to interpolate non-member percentiles.

`quartiles(vals)`
---

Calculate the [quartiles](http://en.wikipedia.org/wiki/Quartile) as an array of three values.

`interquartile(vals)`
---

Calculate the [interquartile](https://en.wikipedia.org/wiki/Interquartile_range) value - `Q3 - Q1`.

`removeOutliers(vals)`
---

Remove the [outliers](https://en.wikipedia.org/wiki/Interquartile_range#Outliers), computed as values outside the interval `Q1 - 1,5 * IQR, Q3 + 1,5 IQR`.



`histogram(vals[, bins])`
---

Expand Down
227 changes: 135 additions & 92 deletions stats.js
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,185 +1,228 @@
"use strict";

module.exports.numbers = numbers
module.exports.sum = sum
module.exports.mean = mean
module.exports.median = median
module.exports.mode = mode
module.exports.variance = variance
module.exports.stdev = stdev
module.exports.percentile = percentile
module.exports.histogram = histogram

var isNumber = require("isnumber")
'use strict';

module.exports.numbers = numbers;
module.exports.sum = sum;
module.exports.mean = mean;
module.exports.median = median;
module.exports.mode = mode;
module.exports.variance = variance;
module.exports.stdev = stdev;
module.exports.percentile = percentile;
module.exports.quartiles = quartiles;
module.exports.interquartile = interquartile;
module.exports.outliersBounds = outliersBounds;
module.exports.removeOutliers = removeOutliers;
module.exports.outliers = outliers;
module.exports.histogram = histogram;

var isNumber = require('isnumber');

function numbers(vals) {
var nums = []
if (vals == null)
return nums
var nums = [];
if (vals == null) return nums;

for (var i = 0; i < vals.length; i++) {
if (isNumber(vals[i]))
nums.push(+vals[i])
if (isNumber(vals[i])) nums.push(+vals[i]);
}
return nums
return nums;
}

function nsort(vals) {
return vals.sort(function numericSort(a, b) { return a - b })
return vals.sort(function numericSort(a, b) {
return a - b;
});
}

function sum(vals) {
vals = numbers(vals)
var total = 0
vals = numbers(vals);
var total = 0;
for (var i = 0; i < vals.length; i++) {
total += vals[i]
total += vals[i];
}
return total
return total;
}

function mean(vals) {
vals = numbers(vals)
if (vals.length === 0) return NaN
return (sum(vals) / vals.length)
vals = numbers(vals);
if (vals.length === 0) return NaN;
return sum(vals) / vals.length;
}

function median(vals) {
vals = numbers(vals)
if (vals.length === 0) return NaN
vals = numbers(vals);
if (vals.length === 0) return NaN;

var half = (vals.length / 2) | 0
var half = (vals.length / 2) | 0;

vals = nsort(vals)
vals = nsort(vals);
if (vals.length % 2) {
// Odd length, true middle element
return vals[half]
}
else {
return vals[half];
} else {
// Even length, average middle two elements
return (vals[half-1] + vals[half]) / 2.0
return (vals[half - 1] + vals[half]) / 2.0;
}
}

// Returns the mode of a unimodal dataset
// If the dataset is multi-modal, returns a Set containing the modes
function mode(vals) {
vals = numbers(vals)
if (vals.length === 0) return NaN
var mode = NaN
var dist = {}
vals = numbers(vals);
if (vals.length === 0) return NaN;
var mode = NaN;
var dist = {};

for (var i = 0; i < vals.length; i++) {
var value = vals[i]
var me = dist[value] || 0
me++
dist[value] = me
var value = vals[i];
var me = dist[value] || 0;
me++;
dist[value] = me;
}

var rank = numbers(Object.keys(dist).sort(function sortMembers(a, b) { return dist[b] - dist[a] }))
mode = rank[0]
var rank = numbers(
Object.keys(dist).sort(function sortMembers(a, b) {
return dist[b] - dist[a];
})
);
mode = rank[0];
if (dist[rank[1]] == dist[mode]) {
// multi-modal
if (rank.length == vals.length) {
// all values are modes
return vals
return vals;
}
var modes = new Set([mode])
var modeCount = dist[mode]
var modes = new Set([mode]);
var modeCount = dist[mode];
for (var i = 1; i < rank.length; i++) {
if (dist[rank[i]] == modeCount) {
modes.add(rank[i])
}
else {
break
modes.add(rank[i]);
} else {
break;
}
}
return modes
return modes;
}
return mode
return mode;
}

// Variance = average squared deviation from mean
function variance(vals) {
vals = numbers(vals)
var avg = mean(vals)
var diffs = []
vals = numbers(vals);
var avg = mean(vals);
var diffs = [];
for (var i = 0; i < vals.length; i++) {
diffs.push(Math.pow((vals[i] - avg), 2))
diffs.push(Math.pow(vals[i] - avg, 2));
}
return mean(diffs)
return mean(diffs);
}

// Standard Deviation = sqrt of variance
function stdev(vals) {
return Math.sqrt(variance(vals))
return Math.sqrt(variance(vals));
}

function interquartile(vals) {
var qs = quartiles(vals);
return qs[2] - qs[0];
}

function outliersBounds(vals) {
var out = [];
var iqr = interquartile(vals);
var qs = quartiles(vals);
var min = qs[0] - 1.5 * iqr;
var max = qs[2] + 1.5 * iqr;

return [min, max];
}

function removeOutliers(vals) {
var minmax = outliersBounds(vals);
return vals.filter(v => v >= minmax[0] && v <= minmax[1]);
}

function outliers(vals) {
var minmax = outliersBounds(vals);
return vals.filter(v => v < minmax[0] || v > minmax[1]);
}

function quartiles(vals) {
return [
percentile(vals, 0.25),
percentile(vals, 0.5),
percentile(vals, 0.75)
];
}

function percentile(vals, ptile) {
vals = numbers(vals)
if (vals.length === 0 || ptile == null || ptile < 0) return NaN
vals = numbers(vals);
if (vals.length === 0 || ptile == null || ptile < 0) return NaN;

// Fudge anything over 100 to 1.0
if (ptile > 1) ptile = 1
vals = nsort(vals)
var i = (vals.length * ptile) - 0.5
if ((i | 0) === i) return vals[i]
if (ptile > 1) ptile = 1;
vals = nsort(vals);
var i = vals.length * ptile - 0.5;
if ((i | 0) === i) return vals[i];
// interpolated percentile -- using Estimation method
var int_part = i | 0
var fract = i - int_part
return (1 - fract) * vals[int_part] + fract * vals[Math.min(int_part + 1, vals.length - 1)]
var int_part = i | 0;
var fract = i - int_part;
return (
(1 - fract) * vals[int_part] +
fract * vals[Math.min(int_part + 1, vals.length - 1)]
);
}

function histogram (vals, bins) {
function histogram(vals, bins) {
if (vals == null) {
return null
return null;
}
vals = nsort(numbers(vals))
vals = nsort(numbers(vals));
if (vals.length === 0) {
return null
return null;
}
if (bins == null) {
// pick bins by simple method: Math.sqrt(n)
bins = Math.sqrt(vals.length)
bins = Math.sqrt(vals.length);
}
bins = Math.round(bins)
bins = Math.round(bins);
if (bins < 1) {
bins = 1
bins = 1;
}

var min = vals[0]
var max = vals[vals.length - 1]
var min = vals[0];
var max = vals[vals.length - 1];
if (min === max) {
// fudge for non-variant data
min = min - 0.5
max = max + 0.5
min = min - 0.5;
max = max + 0.5;
}

var range = (max - min)
var range = max - min;
// make the bins slightly larger by expanding the range about 10%
// this helps with dumb floating point stuff
var binWidth = (range + (range * 0.05)) / bins
var midpoint = (min + max) / 2
var binWidth = (range + range * 0.05) / bins;
var midpoint = (min + max) / 2;
// even bin count, midpoint makes an edge
var leftEdge = midpoint - (binWidth * Math.floor(bins / 2))
var leftEdge = midpoint - binWidth * Math.floor(bins / 2);
if (bins % 2 !== 0) {
// odd bin count, center middle bin on midpoint
var leftEdge = (midpoint - (binWidth / 2)) - (binWidth * Math.floor(bins / 2))
var leftEdge = midpoint - binWidth / 2 - binWidth * Math.floor(bins / 2);
}

var hist = {
values: Array(bins).fill(0),
bins: bins,
binWidth: binWidth,
binLimits: [leftEdge, leftEdge + (binWidth * bins)]
}
binLimits: [leftEdge, leftEdge + binWidth * bins]
};

var binIndex = 0
var binIndex = 0;
for (var i = 0; i < vals.length; i++) {
while (vals[i] > (((binIndex + 1) * binWidth) + leftEdge)) {
binIndex++
while (vals[i] > (binIndex + 1) * binWidth + leftEdge) {
binIndex++;
}
hist.values[binIndex]++
hist.values[binIndex]++;
}

return hist
return hist;
}
Loading