STATISTICA Blogs

Comparing information across multiple plots is a common task. One thing that makes these comparisons easier is common axis scaling. Some three-dimensional plots, such as surface and contour plots, use colors to distinguish levels of the Z axis. This article demonstrates how to achieve a common color coding system for multiple three-dimensional plots.

Of course, categorized graphs in STATISTICA will automatically produce multiple plots, all utilizing the same scale. But, the data may not be structured such that a categorized graph is applicable, or for other reasons it may be more desirable or practical to create individual graphs, while maintaining the same scale across all. This task often requires graph customization. Here, we’ll show how to determine and define the z axis scaling and color scheme.

The technique described will use the three-dimensional contour plot, but is applicable to any STATISTICA graph using surface colors to denote the Z axis. 

The three contour plots below were created using the STATISTICA example data file chemical process.sta. This example file is found in the STATISTICA folder examples > datasets. These plots each show Yield as the Z axis variable, using HCL, H2O, and NH3 as X and Y variables. This example data set does not present a grouping variable, so categorized graphs are not applicable. 

By default, the Z axis color scheme is determined for an individual graph.  When multiple graphs are created, each individual graph is used to determine that graph’s scale.  In this case, that leads to three plots with 3 different color schemes for Z. Dark green is assigned to 57.9 in the first plot, 58.6 in the second and 60.1 in the third. This makes side-by-side comparisons tricky. Each plot uses a different color to denote levels of Z.  

Let’s take a quick look at the same three plots, now using the same Z axis color scheme. The first two are not changed significantly, but the third plot, NH3 by H2O, no longer shows the yellows and green as it did before. The first set of plots may be misleading when careful attention is not given.

How do we achieve the uniform color scheme for the Z axis across multiple plots? The process consists of first determining the axis scaling that will be used, and then applying it to each graph using the graph customization dialogs. The steps are shown below:

1.) Determine the axis scaling to be used. 
Based on the set of existing graphs to be standardized, select a minimum and maximum for the Z axis scaling. In this example, 61 is the maximum used by any graph, and 57.9 is the minimum. These would be appropriate choices.

2.) Modify the first graph. 
Right click on the first graph and select Graph options to display the Graph Options dialog. Under Graph, select Surface.  

Then, click the Surface specs button. In the Surface Specifications dialog, change Get shading levels from to Major tickmarks. Select the Define contours/levels based on min/max of Z-axis check box.

Now click the Vertical (Z) axis tickmarks, options button to display the Vertical Axis dialog. Change the Range Mode to Manual. Specify the new range to Minimum 57.9 and Maximum 61, or your desired Z axis range. This new range will be applied to all graphs.

Click Close in the Vertical Axis dialog and again in the Surface Specifications dialog. Click OK in the Graph Options dialog. The graph will be updated with the new Z axis range.

Repeat this set of customizations for the remaining graphs. 

Graph styles, a tool for saving and applying a set of graph customizations, do not capture these range settings. This is by design. Each individual graph will need to be customized with the desired Z axis range.