replot-boxplot

Intelligent and customizable boxplot components to show off your distributions in style.

Installation

Only works with React projects. React must be installed separately.

npm install replot-boxplot

Then with a module bundler like webpack/browserify that supports CommonJS/ES2015 modules, use as you would anything else.

import BoxPlot from 'replot-boxplot'

API

replot-boxplot is designed to easily create boxplots. The only required input is properly formatted data.

Basic Usage

In the simplest case, just supply data (as a Javascript array) and specify the key for the values -:

render() {
  let boxData: [
    {score: 93},
    {score: 95},
    {score: 60},
    {score: 74},
    {score: 82},
    {score: 90},
    {score: 96},
    {score: 66},
    {score: 74},
    {score: 77},
  ]

  return(
    <BoxPlot data={boxData} weightKey="score" />
  )
}

• data is the only required prop
• weightKey defaults to "weight"

Multiple distributions

Replot BoxPlots support displaying multiple boxplots from one data set. In this case, one must only specify a groupKey prop with which to group the data.

render() {
  let boxData: [
    {gender: "female", score: 93},
    {gender: "female", score: 95},
    {gender: "female", score: 60},
    {gender: "female", score: 74},
    {gender: "female", score: 82},
    {gender: "female", score: 90},
    {gender: "male", score: 96},
    {gender: "male", score: 66},
    {gender: "male", score: 74},
    {gender: "male", score: 77},
    {gender: "male", score: 80}
  ]

  return(
    <BoxPlot data={boxData} weightKey="score" groupKey="gender"/>
  )
}

Outliers Customization

Replot BoxPlots give you the choice in how to draw outliers. By default, outliers are calculated (as min - 1.5IQR or max + 1.5IQR), and drawn as points on the graph, separate from the main box and whiskers.

If the user passes in a showOutliersAsPoints prop with a value of false, then the outliers will be drawn into the BoxPlot graph, not separate points.

Dimensions

Dimensions may be specified by passing in width and height props. The unit is pixels, and the BoxPlot defaults to 400 by 400 pixels.

The BoxPlot will not function with a width that is less than 60 pixels, or with a height that is less than 30 pixels.

Colors

Colors may be specified through 2 different mechanisms, both through a color prop. If none of the mechanisms are specified, BoxPlot defaults to a built in color palette.

User-provided Color Palette

The user can specify their own desired colored palette for the boxplots to use. This is done by passing in an array of color strings to the component with the color prop. The displayed boxplots will cycle through the provided colors.

User-provided Color function

The user can specify the color for a certain distribution by providing a function as well. One can expect to receive the index of the plot (plot on the left has index 0, next plot has index 1, and so on), as well as the groupKey associated with the plot, if there is one. In the example below, color is decided based on the groupKey:

colorMe(i, group) {
  if (group === "male"){
    return "blue"
  } else if (group === "female") {
    return "pink"
  }
}
render() {
  return(
    <BoxPlot data={boxData} weightKey="score" groupKey="gender" color={this.colorMe}/>
  )
}

Tooltip

BoxPlots are capable of utilizing a tooltip to display more specific information about the distribution. By default, the tooltip is off, but can be activated by passing in a tooltip prop (no value needed). The tooltip features two different color schemes, dark and light, which can be specified by a tooltipColor prop, with a value of "dark" or "light".

render() {
  ...

  return(
    <BoxPlot data={boxData} tooltip tooltipColor="light" />
  )
}

Customizing Tooltip contents

By default, the tooltip will display the five-number summary of a distribution when hovering over the respective plot. The user can customize exactly what is displayed inside the tooltip by passing in a tooltipContents prop in the form of a Javascript function. The user can expect to receive the entire distribution as an array, as well as the five (+1) number summary as an object, with keys of max, q3, median, mean, q1, min. The function should return JSX, which can utilize some or all of the provided values.

fillTooltip(distribution, summary){
  return (
    <div>The distribution that makes up this boxplot is {distribution.toString()}</div>
  )
}

render() {
  ...

  return(
    <BoxPlot data={boxData}
      tooltip tooltipContents={this.fillTooltip}/>
  )
}

Graph Style

The BoxPlot offers some customization with regards to the actual graph elements. These can be controlled with a graphStyle prop that is passed in as a javascript object. Keys to include can be the following:

• lineWidth
  • Determines the thickness of the lines drawn on the LineChart
  • Defaults to 2.5
  • Accepts any number value
• fill
  • Determines the fill of the BoxPlot's rectangle
  • Defaults to "rgba(245,245,245,.1)"
  • Accepts any color string

Axis Customization

Replot BoxPlots allow for incredible customization of the graph axis. A complete explanation of axis customization can be found below -:

Titles

By default, the BoxPlot does not display any axis titles. If the user wishes to include titles, they can pass in some or all of the xTitle, yTitle, and graphTitle props. These props accept a string value, displayed at the appropriate location on the graph. To compensate for the inclusion of a title, the graph content will be pushed further in, but overall the size of the component will remain what was specified by the user.

Showing/Hiding Axis Elements

By default, the BoxPlot shows the entirety of the axes, including lines, labels, and gridlines. These can each individually be disabled by passing in boolean values of false to the following props:

• showXAxisLine
• showYAxisLine
• showXLabels
• showYLabels
• showGrid

Styling the axis

In addition to enabling/disabling titles and axis components, the actual style of the components can be altered as well, with the use of one axisStyle prop that takes the form of a JavaScript object.

Explanations and defaults follow:

• axisColor
  • modifies the color of the axis line
  • defaults to #000000
  • accepts any color string
• labelColor
  • modifies the color of both axis labels
  • defaults to #000000
  • accepts any color string
• titleColor
  • modifies the color of all graph titles
  • defaults to #000000
  • accepts any color string
• labelColor
  • modifies the color of axis gridlines
  • defaults to #DDDDDD
  • accepts any color string
• lineWidth
  • modifies the thickness of axis lines
  • defaults to 2
  • accepts any number
• lineOpacity
  • modifies the opacity of axis lines
  • defaults to 1
  • accepts any number

Example of using the axisStyle prop:

let style = {
    axisColor: "#f17e33",
    labelColor: "blue",
    titleColor: "#000000",
    gridColor: "#DDDDDD",
    lineWidth: 5,
    lineOpacity: .5
  }

render() {
  ...

  return(
    <BoxPlot data={boxData} axisStyle={style}/>
  )
}

Initial Animation

Initial animation is enabled by default, resulting in the boxplot growing out from the median of a distribution. This can be disabled using the initialAnimation prop, passing in a value of false.