Ploticus data display engine Copyright Steve Grubb
Ploticus Gallery
Located here are dozens of graphical data
displays, intended to illustrate ways of displaying data
clearly, effectively, and attractively.
These examples also serve to demonstrate the
ploticus data display engine.
Notes on using this gallery
Alphabetical index of all ploticus examples |
All PNG thumbnails |
All GIF thumbnails
The preliminaries to plotting data.
Lineplots are used to show values that change from left to right.
Often the change is over time along the X (horizontal) axis.
Use proc lineplot.
A line plot variant is to fill the area under the curve,
in order to highlight a particular region, to
show a difference between curves, or for stylistic reasons.
Use proc lineplot with the fill option.
Range sweeps may be used to depict ranges that change from left to right.
Often the change is over time along the X (horizontal) axis.
Use proc rangesweep.
Pie graphs may be used to depict proportions that make up a whole, such as
budget categories, or survey breakdowns.
Use proc pie.
Bar graphs (histograms) may be used to show comparisons, distributions, or category
tabulations. They are also sometimes used to show values that
change over time.
Use proc bars.
Horizontal bars are useful for compactness or
where time values are being compared.
Use proc bars with
the horizontalbars option.
Timelines are used to display events and durations with respect to time.
The result may be in the form of a timetable, project progress chart, etc.
Use proc bars with the
horizontalbars and segmentfields attributes.
Bars may be used to show proportions. (A pie graph might also be used).
Use proc bars with the
stackfields option.
Scatter plots display data points in one or two dimensions. Every data
point is plotted with a mark, symbol, or label.
2-D scatterplots are often used to show correlation (or lack thereof)
between two variables.
Use proc scatterplot.
1-D scatterplots show the distribution of one variable.
Lines, symbols, characters or text may be used to mark points.
Use proc scatterplot.
Range bars (box plots)
compactly show the distribution and range of a set of values.
The median (50th percentile) is shown by a dot, the box covers the interquartile
range (25th through 75th percentiles), and the tails show the minima/maxima or the
5th and 95th percentiles.
Use proc rangebar
which can compute the medians, quartiles, etc. as well
as render the bars.
Error bars show the amount of margin of error for a value. This usually is
+/- the standard deviation. If the amount of error has been calculated in
advance, use proc bars. If you want
ploticus to compute the mean and standard deviation, use
proc rangebar with the meanmode option.
Curve fitting may be useful to illustrate
trends in noisy data, or to show where data points lie with
respect to average.
proc curvefit can generate
moving average and bspline curves.
When different symbols, colored bars, colored lines, etc. are
used for comparison, a legend is usually necessary to describe
what the various colors or symbols mean or what they correspond to.
While not the most "sexy" element of a plot, it is often absolutely
necessary. Use the legendlabel attribute of each plotting
proc, then use proc legend to render the legend.
Annotation may be used to denote specific points or regions on a plot
that have some importance, or for any text placement.
Use proc annotate.
Ploticus includes pltab,
which can take ascii tables and lists and render them as is
in a nice font without the use of formatting codes.
Various embellishments may also be done.
The pltab program may also be used to create graphs that are
linked to tables and lists, perhaps overflowing onto multiple pages.