Graph Paneling

Examining Graphs One Group at a Time

When presenting your findings to others, one graph is worth a thousand words. But sometimes a graph can become so crowded that it's difficult to see the big picture. Graph paneling is a simple way to divide a graph into more manageable pieces. It can make the sometimes subtle language of a graph bolder and clearer - without sacrificing data integrity. Here are a few "before and after" examples to show what graph paneling can do.

Compare and Contrast

Before: Relationships within and between the groups in this scatterplot can be difficult to see.

After: Paneling separates the observations by group for easier comparison.

Divide a Chart into 'Bite-Size' Pieces

Before: The sheer number of points (300) in this control chart makes "reading" the graph quite a task.

After: The chart has been split into three panels, each containing a manageable 100 observations.

Organize Data without Worksheet Manipulation

Before: This graph is an overwhelming jumble of distributions.

After: Paneling divides the graph into logical groups without making awkward alterations to the worksheet.

Paneling in Practice

There are three methods for paneling a graph:

• In most cases, these pieces or panels are defined by a categorical variable called a By Variable. A separate panel is created for each value in the By Variable. Access this option from the graph's main dialog box by clicking Multiple Graphs > By Variables.
• With control charts, time series plots, and area graphs, the continuous string of observations is split into panels containing the desired number of equally sized segments. Access these options from the chart's main dialog box by clicking Options > Display.
• After a graph has been created, you can "edit in" panels by choosing Editor > Panel.

Try It Yourself

Work along in Minitab with the following example to see just how easy it is to add paneling to any graph.

The Problem You've decided to plot the relationship between weight and pulse rate. You consider gender and activity level crucial to understanding this relationship, so you include them as grouping variables. The resulting graph is a mess of different symbol types and regression lines. Open the worksheet EXH_TABLE.MTW. Choose Graph > Scatterplot > With Regression and Groups > OK. Under Y variables, enter Weight. Under X variables, enter Pulse. In Categorical variables for grouping, enter Gender and Activity. Click OK.
Edit In Paneling By using one of the categorical variables as a paneling variable, you can alleviate some of the clutter. Choose Editor > Panel. In  By variables with groups in separate panels, enter Activity. Click OK.
Examine Your New Graph The graph is still grouped by gender, but now it's paneled by activity level. As a result, it's easy to compare the effects of activity level and contrast these observations for men and women.

Putting It to Use

As you begin to graph a set of observations, paneling can help you to explore the effects of grouping variables or to "zoom in" on time series data. Later, as you prepare to present your findings to others, paneling can help transform a potentially confusing graph into one your audience can easily understand. Whether you use it to put a magnifying glass to your time series data or to bring order to a tangle of grouping variables, paneling is a powerful way to clarify the things your graph has to say.

Visit Accessing the Power of Minitab for additional tutorials on the many time-saving features and functionality available in Minitab Statistical Software.

Learn more about Minitab 15.

Purchase or upgrade now