@sgratzl/chartjs-chart-boxplot

Chart.js Box and Violin Plot

Chart.js module for charting box and violin plots.

Works only with Chart.js >= 3.0.0

Install

npm install --save chart.js@next @sgratzl/chartjs-chart-boxplot@next

Usage

see Samples on Github

and

Chart

four new types: boxplot, horizontalBoxplot, violin, and horizontalViolin.

Config

The config can be done on a per dataset .data.datasets[0].minStats or for all datasets under the controllers name. e.g., .options.boxplot.datasets.minStats.

interface IBaseOptions {
  /**
   * statistic measure that should be used for computing the minimal data limit
   * @default 'min'
   */
  minStats: 'min' | 'q1' | 'whiskerMin';

  /**
   * statistic measure that should be used for computing the maximal data limit
   * @default 'max'
   */
  maxStats: 'max' | 'q3' | 'whiskerMax';

  /**
   * from the R doc: this determines how far the plot ‘whiskers’ extend out from
   * the box. If coef is positive, the whiskers extend to the most extreme data
   * point which is no more than coef times the length of the box away from the
   * box. A value of zero causes the whiskers to extend to the data extremes
   * @default 1.5
   */
  coef: number;

  /**
   * the method to compute the quantiles.
   *
   * 7, 'quantiles': the type-7 method as used by R 'quantiles' method.
   * 'hinges' and 'fivenum': the method used by R 'boxplot.stats' method.
   * 'linear': the interpolation method 'linear' as used by 'numpy.percentile' function
   * 'lower': the interpolation method 'lower' as used by 'numpy.percentile' function
   * 'higher': the interpolation method 'higher' as used by 'numpy.percentile' function
   * 'nearest': the interpolation method 'nearest' as used by 'numpy.percentile' function
   * 'midpoint': the interpolation method 'midpoint' as used by 'numpy.percentile' function
   * @default 7
   */
  quantiles:
    | 7
    | 'quantiles'
    | 'hinges'
    | 'fivenum'
    | 'linear'
    | 'lower'
    | 'higher'
    | 'nearest'
    | 'midpoint'
    | ((sortedArr: number[]) => { min: number; q1: number; median: number; q3: number; max: number });
}

interface IBoxplotOptions extends IBaseOptions {
  // no extra options
}

interface IViolinOptions extends IBaseOptions {
  /**
   * number of points that should be samples of the KDE
   * @default 100
   */
  points: number;
}

interface IChartJSOptions {
  boxplot: {
    datasets: {};
  };
}

Styling

The boxplot element is called boxandwhiskers. The basic options are from the rectangle element. The violin element is called violin also based on the rectangle element.

interface IBaseStyling {
  /**
   * @default see rectangle
   * @scriptable
   * @indexable
   */
  backgroundColor: string;

  /**
   * @default see rectangle
   * @scriptable
   * @indexable
   */
  borderColor: string;

  /**
   * @default 1
   * @scriptable
   * @indexable
   */
  borderWidth: number;

  /**
   * item style used to render outliers
   * @default circle
   */
  outlierStyle:
    | 'circle'
    | 'triangle'
    | 'rect'
    | 'rectRounded'
    | 'rectRot'
    | 'cross'
    | 'crossRot'
    | 'star'
    | 'line'
    | 'dash';

  /**
   * radius used to render outliers
   * @default 2
   * @scriptable
   * @indexable
   */
  outlierRadius: number;

  /**
   * @default see rectangle.backgroundColor
   * @scriptable
   * @indexable
   */
  outlierBackgroundColor: string;

  /**
   * @default see rectangle.borderColor
   * @scriptable
   * @indexable
   */
  outlierBorderColor: string;
  /**
   * @default 1
   * @scriptable
   * @indexable
   */
  outlierBorderWidth: number;

  /**
   * item style used to render items
   * @default circle
   */
  itemStyle:
    | 'circle'
    | 'triangle'
    | 'rect'
    | 'rectRounded'
    | 'rectRot'
    | 'cross'
    | 'crossRot'
    | 'star'
    | 'line'
    | 'dash';

  /**
   * radius used to render items
   * @default 0 so disabled
   * @scriptable
   * @indexable
   */
  itemRadius: number;

  /**
   * background color for items
   * @default see rectangle.backgroundColor
   * @scriptable
   * @indexable
   */
  itemBackgroundColor: string;

  /**
   * border color for items
   * @default see rectangle.borderColor
   * @scriptable
   * @indexable
   */
  itemBorderColor: string;

  /**
   * border width for items
   * @default 0
   * @scriptable
   * @indexable
   */
  itemBorderColor: number;

  /**
   * padding that is added around the bounding box when computing a mouse hit
   * @default 2
   * @scriptable
   * @indexable
   */
  hitPadding: number;

  /**
   * hit radius for hit test of outliers
   * @default 4
   * @scriptable
   * @indexable
   */
  outlierHitRadius: number;
}

interface IBoxPlotStyling extends IBaseStyling {
  /**
   * separate color for the median line
   * @default 'transparent' takes the current borderColor
   * @scriptable
   * @indexable
   */
  medianColor: string;

  /**
   * color the lower half (median-q3) of the box in a different color
   * @default 'transparent' takes the current borderColor
   * @scriptable
   * @indexable
   */
  lowerBackgroundColor: string;
}

interface IViolinElementStyling extends IBaseStyling {
  // no extras
}

Data structure

Both types support that the data is given as an array of numbers number[]. The statistics will be automatically computed. In addition, specific summary data structures are supported:

interface IBaseItem {
  min: number;
  median: number;
  max: number;
  /**
   * values of the raw items used for rendering jittered background points
   */
  items?: number[];
}

interface IBoxPlotItem extends IBaseItem {
  q1: number;
  q3: number;
  whiskerMin?: number;
  whiskerMax?: number;
  /**
   * list of box plot outlier values
   */
  outliers?: number[];
}

interface IKDESamplePoint {
  /**
   * sample value
   */
  v: number;
  /**
   * sample estimation
   */
  estimate: number;
}

interface IViolinItem extends IBaseItem {
  /**
   * samples of the underlying KDE
   */
  coords: IKDESamplePoint[];
}

Tooltips

In order to simplify the customization of the tooltips the tooltip item given to the tooltip callbacks was improved. The default toString() behavior should be fine in most cases. The tooltip item has the following structure:

interface ITooltipItem {
  label: string; // original
  value: {
    raw: IBoxPlotItem | IViolinItem;
    /**
     * in case an outlier is hovered this will contains its index
     * @default -1
     */
    hoveredOutlierRadius: number;
    /**
     * toString function with a proper default implementation, which is used implicitly
     */
    toString(): string;

    min: string;
    median: string;
    max: string;
    items?: string[];

    //... the formatted version of different attributes IBoxPlotItem or ViolinItem
  };
}

ESM and Tree Shaking

The ESM build of the library supports tree shaking thus having no side effects. As a consequence the chart.js library won't be automatically manipulated nor new controllers automatically registered. One has to manually import and register them.

Variant A:

import Chart from 'chart.js';
import { BoxPlotController } from '@sgratzl/chartjs-chart-boxplot';

// register controller in chart.js and ensure the defaults are set
BoxPlotController.register();
...

new Chart(ctx, {
  type: BoxPlotController.id,
  data: [...],
});

Variant B:

import Chart from 'chart.js';
import { BoxPlotChart } from '@sgratzl/chartjs-chart-boxplot';

new BoxPlotChart(ctx, {
  data: [...],
});

Development Environment

npm i -g yarn
yarn set version 2
yarn
yarn pnpify --sdk

Building

yarn install
yarn build

---

developed by datavisyn.