In LuPaRaMySa/quickchartR: quickcharts.io from R Code!

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

require(quickchartR)

Introduction

Quickchart.io is a powerful application which provides fine and fast charts of different types. The problem is that the way of creating charts is based on a specific URL form, what is not convenient for example big datasets. To improve functionality we have developed quickchartR (pronounce quickcharter), which enables using R language to create Quickchart objects. In quickchartR You can create every type of chart from Quickchart.io except Sparklines ( [See why quickchartR does not support Sparklines](#why-quickchartr-does-not-support-sparklines)):

• Bar Graph - type 'bar'
• Line Graph - type 'line'
• Radar Chart - type 'radar
• Pie Chart - type 'pie'
• Doughnut Chart - type 'doughnut
• Scatter Plot - type 'scatter'
• Bubbel Chart - type 'bubble'

 

 

How to use quickchartR?

---

presumptions

In the following examples we are working with widely known data from Iris and mtcars datasets which are built in R.

Pie chart

We'd like to create a pie chart as simple as possible. This chart will show number of cars with a specific number of cylinders in the database. In order to do so, we need to prepare data:

inputDataPie<-data.frame(table(mtcars$cyl),
  myLabelColumn = rep(c("number od cylinders"),3))
inputDataPie

Then we can use quickchartR() method to plot a chart:

browseURL(
  quickchartR(
    types = "pie",            # type of a chart
    inputData = inputDataPie, # name of dataset (e.g. dataframe)
    xData = cyl,              # X-axis data
    yData = n,                # Y-axis data
    labels = myLabelColumn    #labels
  )
)

Our result:

Doughnut chart

A doughnut chart requires the same data. Let's append some details, for example a title for the chart.

myOptions = list(
  title = list(                        # list of title's options
    display = TRUE,                    # visibility
    text = "My first quickchart",      #text
    fontSize = 32 
  ),
  legend = list(position = "bottom"))  #possible values: botton, top, left, right, false

As You see type myOptions is list which contains another lists. The result is shown below:

browseURL(
  quickchartR(
    types = "doughnut",
    inputData = inputDataPie,
    xData = cyl,
    yData = n,
    labels = myLabelColumn,
    options = myOptions,
    base64 = F
  )
)

Please note we add "base64" argument as FALSE which causes that a hyperlink to the chart is not encoded.

Bar chart

Another useful type of chart is barchart. Ours will show difference of mean sepal length for three iris species from iris dataset. To achieve that we need to prepare data:

data <- iris
head(data)

mean_setosa_sepal_length<-mean(data[data$Species=="setosa",1])
mean_virginica_sepal_length<-mean(data[data$Species=="virginica",1])
mean_versicolor_sepal_length<-mean(data[data$Species=="versicolor",1])
mean_vector<-c(mean_setosa_sepal_length,mean_virginica_sepal_length,mean_versicolor_sepal_length)
mean_vector
inputData2 = data.frame(
  myXColumn =c("setosa","virginica","versicolor"),
  myYColumn = mean_vector,
  myLabelColumn = rep("Iris", 3)
)

myOptionsb = list(
  title = list(                       
    display = TRUE,               
    text = "Mean of sepal length for different types of iris", 
    fontSize = 32 
  ))

browseURL(
  quickchartR(
    types = "bar",
    inputData = inputData2,
    xData = myXColumn,
    yData = myYColumn,
    labels = myLabelColumn,
    options=myOptionsb
  )
)

Radar chart

Note: There are some issues regarding using radarchart in Google Chrome.

For setosa species we will show all mean dimensions for petals and sepals. In addition we add our own color pallete.

dt<-data.frame("setosa",t(colSums(data[data$Species=='setosa',1:4]))/length(data[which(data$Species=="setosa"),1]))
inputData3 = data.frame(
  myXColumn =c("Sepal Length","Sepal width","Petal Length","Petal width"),
  myYColumn = as.vector(t(dt[,-1])),
  # = c(1,2,3,4),
  myLabelColumn = c(rep("setosa", 4))
)

myOptionsr = list(
  title = list(                       
    display = TRUE,               
    text = "Means of sepal and petal dimensions of iris", 
    fontSize = 32 
  ),
  legend = list(position = "bottom"))

detailedOptions = list(list(fill='true'))
browseURL(
  quickchartR(
    types = "radar",
    inputData = inputData3,
    xData = myXColumn,
    yData = myYColumn,
    #rData = myRColumn,
    labels = myLabelColumn,
    options=myOptionsr,
    detailedOptions=detailedOptions,
    colors = c("blue","red"),
    base64 = F
  )
)

Please note that a radarchart could be misleading if you use incomparable types of data (for example a horse power for an engine and dimensions of a car.)

Scatter plot

We can also create a basic scatter plot:

inputDataScatter = data.frame(
  myYColumn =mtcars$hp,
  myXColumn = mtcars$qsec,
  myLabelColumn = rep(c(" "),length(mtcars[,1]))
)

myOptionss = list(
  title = list(                       
    display = TRUE,               
    text = "Horse power versus 1/4 mile time", 
    fontSize = 32 
  ))

browseURL(
  quickchartR(
    types = "scatter",
    inputData = inputDataScatter,
    xData = myXColumn,
    yData = myYColumn,
    rData = myRColumn,
    labels = myLabelColumn,
    options=myOptionss,
    base64 = F
  )
)

Bubble plot

A bubble plot is a variant of a scatter plot, but it contains additional information represented by a size of the bubble.

inputDataBubble = data.frame(
  myYColumn =mtcars$hp,
  myXColumn = mtcars$qsec,
  myRColumn = mtcars$mpg,
  myLabelColumn = rep(c(" "),length(mtcars[,1]))
)

myOptionsbb = list(
  title = list(                       
    display = TRUE,               
    text = "Horse power versus 1/4 mile time versus mile per gallon distance", 
    fontSize = 22 
  ))

browseURL(
  quickchartR(
    types = "bubble",
    inputData = inputDataBubble,
    xData = myXColumn,
    yData = myYColumn,
    rData = myRColumn,
    labels = myLabelColumn,
    options = myOptionsbb,
    base64=F
  )
)

Mixed charts

QuickchartR allows also to present different charts in one space, called a mixed chart. Data are grouped by their labels and a types argument becomes a vector of strings.

types = c("bar", "line", "bar")
detailedOptions = list(list(), list(fill = "false"), list())
inputDataMix = data.frame(
  myXColumn = rep(1:6, 3),
  myYColumn = sample(1:18, 18),
  myLabelColumn = rep(letters[1:3], 6)
)

browseURL(
  quickchartR(
    types = types,
    inputData = inputDataMix,
    xData = myXColumn,
    yData = myYColumn,
    labels = myLabelColumn,
    detailedOptions = detailedOptions,
    base64 = F
  )
)

 

 

FAQ

why quickchartR does not support sparklines

In its assumption of "simplicity" which ignores labels, axis etc. sparkline chartit is considered a terrible data visualization practice. Even though it is usually used e.g. inside a table to minimize the eye-candy and with a common axis with other charts, implementing it here would cause someone who is not familiar with advanced data visualization and good practices to make a mistake (e.g. by using it regardless of its true purpose). It is assumed that professional data scientists do NOT use Quickcharts for their charts because they have their own advanced tools for it. Therefore, to somehow control and help the newbie data visualizators, sparkline chart is NOT included in quickchartR.

LuPaRaMySa/quickchartR documentation built on Jan. 23, 2020, 3:49 a.m.