To HomePage

R Graphics

Scatterplots

• Use the cars data frame that installs with R.
   
• Use help(cars) to get information about this dataset from 1920. It shows the stopping distance for a car (ft) when driving at a given speed (mph).
   
• Then type
  

  plot(cars)
  

  to produce this scatterplot.  The input parameter to plot, when used with one parameter, must be a data frame or matrix with two columns.
   
• A more common way to use the plot function is
  

  plot(x, y)
  

  where x is the independent variable vector and y is the dependent variable vector. We can use the plot function with two input parameters like this:
  

  x <- cars[,1]
  y <- cars[,2]
  plot(x, y)
  

• To produce a more informative graph, here are some plot function parameters that can be used:
   
  

  Parameter	Description
  main	Title of the plot at top of graph
  sub	Subtitle of the plot at the    bottom of graph
  xlab	Label for x-axis
  ylab	Label for y-axis
  asp	Aspect ratio of data ranges
  xlim	Mininum and maximum of x-axis range
  ylim	Mininum and maximum of y-axis range
  pch	Plotting symbol (character or numeric value from 1 to 25)
  col	Color of plotting symbol
  type	Type of plot1

  
  
• ¹Possible plot types:
   
  
  "p": points (default)
   
  "l": lines
   
  "b": both
   
  "c": lines only of "b"
   
  "o": overplot lines over points
   
  "h": high density vertical lines
   
  "s": stair steps, first horizontal, then vertical
   
  "S": stair steps, first vertical, then horizontal
   
  "n": no plotting; Set up plotting region with labels and titles only.
  
  
• Here is a script that plots the Cars data using some of the common R graphics options:
  

  x = cars[ ,1]
  y = cars[ ,2]
  plot(x, y, main="Cars Dataset",
     sub="Data from 1920s",
     xlab="Speed (mph)",
     ylab="Stopping Distance (feet)",
     xlim=c(0, 30), ylim=c(0, 130),
     pch="*", col="red")
  

• See the Cars1 Example, which shows nine plots of each of the possible plot types.
   
• The Cars2 Example shows the points plotted on a logarithmic scale.
   
• The loess smoother uses local weighted polynomial regression to fit a nonparametric curve to data.

Redirecting Graphics Output to Disk

Setting Graphical Parameters

• Although some graphical parameters can be set in the plot function, such as col and pch, some graphical parameters must be set using the par function. Others can be set in either place. Here is a list of the parameters that can be set using the par function. Parameters marked with * must be set with the par function.
  

  adj  ann  *ask  *bg  bty  cex  cex.axis
  
  cex.lab  cex.main  cex.sub  cin  col  
  
  col.axis  col.lab  col.main  col.sub
  
  cra  cxy  family  fg  *fig  *fin  font
  
  font.axis  font.lab  font.main  font.sub  
  
  lab  las  lend  *lheight  ljoin  lty
  
  lwd  *mai  *mar  *mex  *mfcol  *mfrow  
  
  mgp  *new  *oma  *omd  *omi  pch  *pin
  
  *plt  *ps  *pty  srt  tck  tcl  usr  xaxp  
  
  xaxs  xaxt  *xlog  xpd  xaxp  yaxs  yaxt  
  
  *ylbias  *ylog
  

• See this R Graphics Parameters table or help(par) to see details about these parameters.
   
• Type par( )$pch or par("pch") to see current settings of the par pch parameter. (Substitute desired parameter for pch.)

Setting Colors

• Colors for R graphics can be specified in four ways
   
  
  1. By name. Example:
     

     col="DarkKhaki"
     

  2. By hex color coordinates
     

     col="#808000"
     

  3. From one of the R predefined color palettes: rainbow, heat.colors, terrain.colors, topo.colors, cm.colors. For example:
     

     col=heat.colors(4)
     

     uses the first four colors of the heat.colors palette "#FF0000FF", "#FF8000FF", "#FFFF00FF", "#FFFF80FF". (The fourth component of each color is the alpha component, which is the transparency.
      
  4. By setting a defined color pallet:
     

     palette(c("red", "green", "blue"))
     

• See these examples:
   
  
  AutoSales1   AutoSales2

More Charts and Graphs

• Bar Plots   Use the barplot function. Important parameters are:
  

  height  Width  space  names.arg  legend.text
  
  beside  horiz  density  angle  col  border
  
  main  sub  xlab  ylab  xpd  log  cex.axis  
  
  cex.names  plot

  See the AutoSales3 Example.
   
• Pie Charts   Use the pie function. Important parameters are:
  

  x  labels  edges  radius  init.angle  density
  
  angle  col  border  lty  main
  

  See the PieCharts Example.
   
• Dot Plots   Use the dotplot function. Important parameters are:
  

  x  labels  groups  gdata  cex  pch  gpch  bg
  
  color  gcolor  lcolor  xlim  main  xlab  ylab
  

  See the DotCharts Example.
   
• Histograms   Use the hist function. Important parameters are:
  

  x  breaks  freq  include.lowest  right  
  
  density  angle  col  border  main  xlab  
  
  ylab  xlim  ylim  axes  plot  labels    
  

• The breaks parameter can be specified in these ways:
   
  
  1. a vector giving the breakpoints between histogram cells,
      
  2. a single number giving the number of cells for the histogram (suggestion only),
      
  3. a character string naming an algorithm to compute the number of cells.  The choices are "Sturges", "Scott", "FD" (Friedman-Diaconis),
      
  4. a function to compute the number of cells.
  
  
• If x is the data vector and the bin width is h, then the number of bins k can be calculated as
   
  
  k = ⌈(max x - min x) / h⌉
  
  
• The breaks parameter can accept one of these three algorithms for automatically choosing the number of histogram bins.
   
1. Sturges's Rule   (breaks="Sturges")
    
   
   h = ⌈1 + log₂n⌉
   
   
2. Scott's Rule   (breaks="Scott")
    
   
   h = 3.5 s_x / n^1/3
   
   
3. Freedman-Diaconis Rule   (breaks="FD")
    
   
   h = 2 * IQR / n^1/3
   
   
4. Create a histogram from these simulated IQ scores:
   

   scores = rnorm(85, mean=100, sd=15) 
   

• Boxplots   Use the boxplot function. Important parameters are:
  

  formula  data  subset  x  range  width
  
  varwidth  notch  outline  names  boxes 
  
  plot  border  col  log  horizontal  add
  
  at  
  

• See the Boxplots Example.