Welcome to JMP 15: It’s got instant graphical gratification and more
Oct 16, 2019 4:09 PM
| Last Modified: Oct 16, 2019 4:18 PM
John Sall, chief architect of JMP, unveils the newest version of the software at Discovery Summit Tucson 2019.
The opening plenary session of Discovery Summit Tucson 2019 featured John Sall, chief architect of JMP, unveiling the latest version of JMP statistical discovery software from SAS. Sall officially announced the release of JMP 15, guided the audience down the first 30 years of JMP memory lane through the new release, and introduced JMP development directors who presented enhanced and brand-new JMP features.
Not only is JMP 15 now available, but also it comes as JMP celebrates its 30th birthday. Sall described three big steps in the evolution of JMP:
1989: JMP 1 — Emergence of the graphical user interface increased the importance of statistical graphics and growth into the enterprise space.
1995: JMP 3.2 — Standardization of JMP on Windows after near-death of Mac and the emergence of the Six Sigma movement.
2000: JMP 4 — Rewrite of JMP, including shift to C++, truly portable layering, modern DOE, and scripting.
Here we are at 2019 and JMP 15.
JMP development directors Brian Corcoran, Xan Gregg, Chris Gotwalt, and Distinguished Research Fellow Bradley Jones presented brief demos of key new features in JMP 15.
Brian Corcoran ( @briancorcoran ) demonstrated new and enhanced hosting features:
JMP Help is all on the web now for easier content updates between releases and a smaller JMP install footprint.
Improved crash reporting means a simpler interface for submitting reports right when a crash happens.
Enhanced HTTP Request offers easier retrieval of data from web providers, for example, importing the location of breweries in Arizona into JMP for analysis (quite appropriate for a Discovery Summit event in Tucson).
Xan Gregg ( @XanGregg) presented the ways JMP 15 brings instant graphical gratification — new Graph Builder functionality and the many new places to look for graphs in JMP:
In Graph Builder, there is a new time series line of fit; box plots and heatmaps have new coloring options; bagplots (2D box plot) are now in the Contour element; histograms have more options for y-axis labeling; and new ridgeline plots are available.
There are new ways to get data into JMP — the JSON import and XML import have new wizard interfaces, and data from PDF files can now be accessed using the new PDF import wizard.
Gregg quoted John Tukey, “There is no excuse for failing to plot and look.” In JMP, that is true even more so. There are new histograms and bar charts in the local data filtering controls, graphlets that appear when hovering over points in certain platforms (for example, in Principal Components Analysis) and header graphs above columns in the data table for preliminary data exploration and quality checking (which drew an excited gasp from the audience).
Bradley Jones ( @bradleyjones ) described the new Group Orthogonal Supersaturated Designs platform for creating and analyzing supersaturated designs with groups of correlated effects. The platform offers an easy button for creating these designs, and provides both designer and fitter options. For the details, read the full Technometrics paper, Construction, Properties, and Analysis of Group Orthogonal Supersaturated Designs.
Chris Gotwalt ( @chris_gotwalt1 ) walked through major enhancements to the analytical capabilities of JMP that elicited some excited murmurings throughout the audience. Time series forecasting offers automated options for using time series data to predict into the future, new Model-driven Multivariate Control Charts give an overall summary of process health across many columns, and the newly embedded Generalized Regression platform within Functional Data Explorer makes functional response DOE possible. Additionally, two new platforms join the JMP Pro modeling toolkit: Support Vector Machines (SVM) and Structural Equations Models (SEM).
Finally, John Sall ( @john_sall ) took the stage to demonstrate two of his favorite features in JMP 15:
The new Explore Patterns platform offers another way to know what is in the data before analyzing it. The platform lets users uncover any potential data manipulation, bad mechanisms in a process or falsification of data.
Enhancements to the contour profiler provides a look at multiple contour plots simultaneously.
The latest functionality available in JMP software creates new possibilities for how users access, explore, graph, model and share data. Welcome to JMP 15!