diggingnumbers:pictorial_displays

Pictorial displays are among the most important techniques that help you describing and analyzing your data. The ** R** graphics system is very powerful and lets you produce professional-looking graphics. There are high-level plotting functions that are best suited for simple graphs, while low-level functions provide you with advanced tools to edit details and add annotations.

These examples are based on the high-level plotting functions. Most of the times getting the right result is a matter of playing with some of the many graphical parameters. General graphical options are handled by the `par()`

function, while each specific function has its own parameters.

Barcharts are good if you have to depict a non-numeric variable, such as nominal and ordinal variables. Bars in the graph are physically **separated** meaning that this is not a cartesian graph.

> barplot(table(Cond))

> barplot(table(Mat,Cond), beside=FALSE)

> barplot(table(Mat,Cond), beside=TRUE)

Histograms are as much different from barcharts as ratio variables differ from nominal. An histogram represents the density distribution of values on a continuous axis.

> hist(Socle, freq=TRUE)

These are known also as “box-and-whisker plots”, because of their shape. Boxplots are useful to compare visually the same variable in different datasets because they provide a quick way to represent the most common measures of position (mean, quartiles, outliers).

> boxplot(Socle)

In this graph the central line in the box represents the average mean value, the alone point is an outlier.

Stem-and-leaf plots are useful if you need not only to represent the distribution of your variable, but also to keep your original data available without losing too much space.

> stem(Socle) The decimal point is at the | 2 | 40114455 4 | 23556812489 6 | 01466258 8 | 011466726 10 | 2 12 | 5 14 | 4

> stem(Socle, scale=2) The decimal point is at the | 2 | 4 3 | 0114455 4 | 235568 5 | 12489 6 | 01466 7 | 258 8 | 0114667 9 | 26 10 | 2 11 | 12 | 13 | 5 14 | 4

Scatterplots are a mean to compare one variable against another, plotting on a cartesian surface. The values of one variable are used as X values, and the other's as Y values.

> plot(Maxwi,Maxle, col=Mat, pch=19)

> plot(Maxwi,Maxle, col=Mat, pch=Mat)

Start · Data description · Transforming variables · Tables · **Pictorial displays** · Measures of position and variability · Sampling · Tests of difference · Tests of distribution · Correlation · Tests of association

diggingnumbers/pictorial_displays.txt · Last modified: 2012/11/30 09:58 (external edit)