Fundamental ODS Graphics: Axis Alternatives. Sanjay has wowed me personally with a few associated with the graphs which he has established recently in Graphically Speaking.

In this website, we will revisit a number of his graphs (and some of mine) while focusing on just the axes, grid lines, and guide lines. They may never be the essential exciting areas of a graph, but you can find numerous choices that whenever correctly used are able to turn a great graph into a graph that is great. The axis options consist of most of the choices that one can set when you look at the XAXIS, X2AXIS, YAXIS and Y2AXIS statements (such as axes lines, ticks, tick labels, axis labels, and grid lines).

Why don’t we start out with a fundamental, default pair of axes. We’ll utilize certainly one of my graphs from Vector Plots and Point that is adjusting Labels. By standard, PROC SGPLOT produces axes that are horizontal the underside (X) and top (X2) and straight axes regarding the left (Y) and right (Y2). The X and Y axes have a label for every axis and marks and labels for every tick. By standard, there are not any grid lines or guide lines.

Certainly one of my objectives to make this graph would be to reduce the vectors, which by standard emanate through the beginning (0,0). A short vector combined with familiarity with the foundation is enough to demonstrate the way of this vector. Eliminating ink that is”redundant makes a graph this is certainly cleaner compared to the standard. I shall say more info on Tufte’s idea of “data ink” following the graph that is next.

I first looked at composing a post on axes once I had been considering Sanjay’s Image Backgrounds weblog. This graph is striking due to the back ground, which means you may not really take into account the axes. But, the decision of specifying choices that suppress most of the standard the different parts of the axes is crucial for the look of this graph.

Sanjay utilizes these statements:

The X axis does not have any label, no axis line, with no ticks; this has only tick labels. The Y axis doesn’t have label, no axis line, with no ticks; it has tick labels and grid lines. The sole lines would be the horizontal grid lines. Neither axis requires a label. That is real for date axes along with other axes where in actuality the tick labels require no explanation that is further. Tick labels may not require description because they’re fully descriptive as it is (age.g. times) or as they are explained somewhere else when you look at the graph such as for instance in a title or footnote. Axis labels, axis lines, and tick markings could have cluttered this graph.

Edward Tufte popularized the notion of information ink–the part that is essential of graph. Tufte makes an instance for making the most of information ink while minimizing anything else including redundant ink and “chartjunk”. No body would explain a graph with a background image of a sky and a rainbow as minimizing nondata ink. Nonetheless, eliminating the axis lines and labels do eliminate some ink that is nondata. My objective just isn’t to argue for or against making the most of information ink or minimizing chartjunk—personally, i do believe Sanjay’s usage of images, outlines, and information skins make those graphs more striking and engaging—I simply want to explain to you that one can earn some stunning graphs to some extent by maybe not showing most of the axis elements.

You can observe axis that is similar in Lollipop Charts. This time around there was a horizontal axis line into the graph that presents vertical lollipops but no axis lines within the graph that shows horizontal lollipops. Both look good. You can see axis that is similar in Scatter with suggest Value.

Next, why don’t we have a look at Clinical Graphs: Waterfall Plot ++. The graph that is first no horizontal axis, however it has the full straight axis that includes a line, label, ticks, and tick labels. The next graph is comparable, however it contains two maps and a Y and Y2 axis.

Multipage undesirable Event Reports displays an X2 axis with a line and tick labels. The X axis is made from only a relative line, and absolutely nothing is exhibited regarding the Y axes. The horizontal that is additional are guide lines. Right here the target is to make a multipage report that combines tabular and graphical information.

See Animal Life Expectancy Graph for a typical example of a graph which have no axes. A TEXT declaration provides tick label information not when you look at the old-fashioned tick label places. Also see A Graph with artistic Categories for any other types of minimal axis options combined with the utilization of pictures.

Glance at other Graphically posts that are speaking other samples of axes. The Stem and Leaf Plot post produces a graph that includes no axis that is horizontal. The Y axis has an axis line and tick labels but no ticks and no axis label. Diagonal Tick Values shows dealing with long tick labels regarding the horizontal axis. The SAS/STAT documents chapter ODS Graphics Template Modification has examples that are additional. The top and right axis lines, and display only reference lines and no axes in particular, Example 22.4 Customizing Axes and Reference Lines shows how to put reference lines through the middle of a graph, suppress. That area centers on GTL rather than PROC SGPLOT, you could effortlessly utilize PROC SGPLOT to complete the things that are same.

Observe that despite having NOLINE specified into the two axis statements, you nonetheless still need to specify NOBORDER when you look at the PROC statement to suppress the axis fully lines.

To sum up, you will find four statements that control axes: XAXIS, X2AXIS, YAXIS, and Y2AXIS. Each has a DISPLAY= option that allows one to specify DISPLAY=ALL (the standard), DISPLAY=NONE, or DISPLAY=(NOLINE NOTICKS NOLABEL NOVALUES) and select which section of each axis to suppress. There are numerous other available choices like the GRID option, which displays grid lines at each tick. You may show guide lines. Start to see the XAXIS Statement and REFLINE Statement to learn more. Managing these statements and options allows you to make clean, elegant graphs.

About Writer

Warren F. Kuhfeld is a research that is distinguished designer in SAS/STAT R&D. He received their PhD in psychometrics from UNC Chapel Hill in 1985 and joined up with SAS in 1987. He has got utilized SAS since 1979 and has now developed SAS procedures since 1984. Warren composed the SAS/STAT paperwork chapters “Making use of the Output Delivery System,” “Statistical Graphics utilizing ODS,” “ODS Graphics Template Modification,” and “Customizing the Kaplan-Meier Survival Plot.” He additionally penned the free internet publications fundamental ODS Graphics Examples and Advanced ODS Graphics Examples.