Designed for the technician, scientist, engineer and businessperson
It should come as no surprise to readers of this column that JMP is a personal favorite and, along with SAS, one of my most-used programs. There are a number of reasons for this. Of the many advantages that most packages can offer, breadth and depth of the statistics offered, quality of the diagnostics, interconnectivity of graphics with both data and analyses, and ease-of-use issues are uppermost in my mind as most desirable. A close second is ease of navigation though the main screen and ease of correctly formatted data importation. JMP excels in every category. While SAS is made for the statistician, JMP is designed for the technician, scientist, engineer and businessperson.
First, let’s take a look at the main screen (Figure 1). From this, we can navigate to a previously used file (upper left area), choose a recently used help area (lower left area), or open a previously saved file instantly (right half of navigation area). In addition, you have access to all of the JMP functions from both the main toolbar and the icon toolbars for instant connect to most main or sub-functions.
JMP has long been known for the excellence of its graphics, its ease-of-use, and connecting a graphic to almost any analysis. With each new version, the ease-of-use is enhanced, as well as newer capabilities such as the graph builder, the table builder (called Tabulate) and the custom profiler. In addition, the DOE (experimental design) menu has had the newest space filling model and definitive screening designs included. New strategies for comparing screening designs are also quickly implemented by correlation cell plots. The following is a brief overview of the new features:
• Definitive Screening Designs
• EXCEL Import Wizard
• Transform Variables
• Consumer and Market Research
• Response Screening
• Street-Level Maps
• Assess Variable Importance
• Smarter Filtering and Summaries
• Save Interactive HTML Reports with Data
• Build Attribute and Rare Event Charts
• Crisper, Cleaner Design
• Import Sampling for SAS Data and Text Files
• MATLAB Integration
• JSL Development Environment
• Multiple Comparisons Made Easy
Before walking through an example illustrating both JMP’s ease-of-use and powerful analytic capabilities, a mention must be made of JMP’s extremely useful “Preferences” option. In this area, the user may select only those graphic and diagnostic outputs that they really want to see. JMP will present a large number of outputs for almost any analysis, and this section will limit the output to only those most pertinent to the particular analysis. Thus, the outputs are customized to the individual. These options are available for most platforms. Now, on to a specific example…
Principal Component analysis (PCA) is a sophisticated technique for doing group separations and predictive modeling and exploratory data analysis. As the JMP manual states: “The purpose of principal component analysis is to derive a small number of independent linear combinations (principal components) of a set of measured variables that capture as much of the variability in the original variables as possible. A principal component analysis models the variation in a set of variables in terms of a smaller number of independent linear combinations (principal components) of those variables.”
For our example, we may use the JMP data set solubility.jmp. This quantitates the solubility of a series of organic compounds in a set of seven standard solvents. We simply select Analyze/Multivariate Methods/Principle Components, and the dialog box appears (Figure 2).
The seven compounds have been selected and entered by clicking and dragging them to the Y, Columns box. Upon pushing the <OK> button, the summary plots appear (Figure 3).
The online JMP help section on PCA tells us: “The report gives the eigenvalues and a bar chart of the percent of the variation accounted for by each principal component. There is a Score Plot and a Loadings Plot as well.
The eigenvalues indicate the total number of components extracted based on the amount of variance contributed by each component.
The Score Plot graphs each component’s calculated values in relation to the other adjusting each value for the mean and standard deviation.
The Loadings Plot graphs the unrotated loading matrix between the variables and the components. The closer the value is to 1, the greater the effect of the component on the variable.”
This isn’t all, however, as we can click on the red arrow (hot spot) at the top bar labeled ‘Principal Components on Correlations, to get all of the options (Figure 4).
Of all of the options, I have found the Scree plot and Scatterplot 3D to be the most useful (Figures 5 and 6).
The break in the Scree curve at ‘2’ or ‘3’ tells us the optimal number of Principal Components that are needed for optimal separation of groups or variables. The 3D plot visualizes this for us. As was seen above, there are many more options.
JMP has many advantages as both an analysis and graphics software platform. It is extensively used for science and business applications, as well as engineering and technology. It is recommended that the reader download a free 30-day evaluation copy and see for themselves. Highly Recommended.
• $1,470 single license, commercial, perpetual
SAS Campus Drive, Building T.
Cary, NC 27513; 877-594-6567
John Wass is a statistician based in Chicago, IL. He may be reached at editor@ScientificComputing.com.