New in JMP^® 19

Data access, prep, and exploration

Graph Builder updates

• Optimized marker jitter adds more options to dot plots.
• Live view for Gradient settings has been added.
• Smoother Constraints support shape constraints on the result of the smoother, such as being monotonic.
• Parallel Y Axes option allows many independent Y axes to be on the left and right side of a graph.
• Comfy transforms allows summarizing calculations to be customized. When making transform columns that depend on multiple rows, comfy transforms provides a way to control how those rows correspond to platform context.
• Vector fields have a new interval style, Arrow, which creates a vector field view that consists of many small arrows.
• Functions that support transform scope have been added.
• Learn more about Graph Builder updates.

Additional items

• Categorical data filter can search for multiple selections and not clear the previous selections.
• Expression columns with image file types .png or .jpeg are exported to Microsoft Excel in Export/Save As to Excel.
• Tabulate can use multiple response columns in the drop zones and as a Page Column variable.
• Tabulate provides a better way to format tables within the dialog.
• The tabulate dialog now provides a more consistent experience when using larger tables or tables with many columns, as well as when making multiple selections in the control panel.
• Oracle and Teradata schema and table retrieval are more responsive.

Design of experiments, reliability, and process and quality engineering

Control Chart Builder updates

• Sort By Subgroup is no longer turned on by default in Control Chart Builder. This changes the default order of the horizontal axis when there is a subgroup variable.
• The Last N Subgroups option that displays a rolling window of the last N subgroups.
• Option to set the moving range span.
• Dispersion Chart options in IMR chart dialog.
• Connect thru Missing for connecting through excluded and hidden.
• Option to view only the most recent 'n' observations.
• Preference to add Legacy Control Charts back to menu.
• Ability to create a run chart with more than one observation per subgroup.
• Ability to alter KSigma in launch window.
• Constant sample size option in attribute chart launch window.
• New launch window options for Get Limits and Capability.
• Enhanced OC curves for easier customization.
• P charts for proportion data.
• Additional options in IMR on Means launch window.

Constant Stress Accelerated Life Test (CSALT) update

The new model simulator for CSALT runs simulated experiments based on one of the generated designs and an assumed model.

Definitive Screening Design update

Definitive Screening Design now allows additional extra runs and supports the addition of center points for designs with no blocks or categorical factors.

Design Explorer update

Create a definitive screening design (DSD) with varying numbers of extra runs and center points to compare to other DSDs or custom designs.

Mixed-level screening designs with categorical factors

Perform mixed-level screening with a mix of continuous factors and categorical factors that can have more than two levels.

Process Capability enhancements

Updates include an option to add group-by columns; normal probability plot to assess whether normal distribution is a good fit; LEV and SEV distributions; preference for spec limit line colors; additional statistics in the interactive capability plot; and Save In Spec (or Out of Spec) indictor column option in the Save menu.

Sample size explorers updates

New sample size explorer for nonparametric tolerance interval confidence explorer, and new power explorer for two-paired sample means have been added. Many of the explorers now use the functionality of the standard profiler.

Additional items

• Bayesian fault localization (BayesFLo) in DOE Covering Array platform introduces a probabilistic ranking of suspicious combinations that allows you to specify the suspicion level of different factors based on domain knowledge and incorporates such knowledge into fault localization process. (JMP Pro)

• Degradation Profiler has been changed to Degradation Quantile Profiler to include more information in the analysis.

• Enhanced Simulate Responses extends Random Coefficients to most DOE platforms and allows selection of whether to enforce effect sparsity, effect hierarchy, and effect heredity.

• Enhancements to OC curves add a target parameter to all OC curves and more descriptive text for the Xbar OC curve.

• Fatigue Model updates include a plot of standardized residuals, a new profiler that enables you to explore nonzero failure probability at infinity, a multiple probability plot, and two new fatigue-strength models.

• Manage Limits allows you to add Group-by Columns to specify process metadata by grouping variables.

• Recurrence improvements to modeling when there are a large number of causes. Updates include:

  • MCF Differences support for multiple causes.
  • Improved handling of singularities.
  • New command to Save MCF Differences.
  • New menu item, Show MCF Legend, for hundreds of causes or groups.
  • Fitting results in report tables.
  • Model summary with a frequency count for a cause if it has a Grouping column in the model.
  • More clarity on cause (failure mode versus causal modeling).

• Reliability Growth updated to change all by-system graphics from stacking graphics to overlay graphics.

• Reliability menu reorganized for easier use.

• Type 1 Gauge Summary and Capability Statistics report updated to include Hysteresis and Drift, as well as the option to customize summary statistics.

• The Variance Inflation Factor (VIF) added to Evaluate Design as a hidden column.

• An option to specify the number of whole plots in Augment Design.

• Enhanced capabilities in Parametric Survival:

  • Likelihood confidence intervals for estimating probability and quantile.
  • Residual vs. fitted median and response vs. fitted median plots.
  • Survival function profiler.

• Learn more about DOE updates.

Statistics and modeling

Design Space Profiler updates

Find the worst in-spec probability in the current space using the new option, Show Corners. Includes buttons to move inward and outward from cold corners and cube plots in the Corners report.

Distribution update

Two new distributions – SEV and LEV – have been added to the list of distribution fitters. There are now new nonnormal tolerance intervals, in addition to existing normal and nonparametric tolerance intervals options. Support for limits of detection in fitting discrete distributions and new censoring ability are now also available.

Fit Model B-Spline effects

B-Splines support general continuous smooth curve-fitting and can specify the degree of the spline, the number of internal knot points, and the positions of the knot points.

Sliced effect comparisons

Create multiple comparison reports for slices of interaction effects in Fit Least Squares, REML, Fit Mixed, Fit GLMM, and Generalized Regression.

One-way analysis

New tests have been added to the Compare Means menu. The Games-Howell Post Hoc test supports multiple comparisons on means assuming unequal variances; Fieller's Method conducts a means ratio comparison using either pooled variance or unequal variance.

Profiler features

New features in profilers include:

• Option to import the linear constraints from a data table to the profiler.
• Option to load from and save to linear constraint data tables in the profilers.
• Clipping region to show contours for wafer maps in Contour Profiler.
• Greater flexibility for Simulation Experiment, including supporting fixed factors.
• New options in the Simulation Experiment window to make the platform more automatic.
• Reorganized profiler menus.
• Set Desirabilities dialog’s ability to handle many responses in one window.
• New Optimization control panel.
• New filled contour option for the contour profiler.
• Data points that can appear faded and smaller in size, based on their distance from the plane of each profiler.

Response Screening enhancements

Response Screening functionality has been updated to include:

• Subgrouping.
• Fitting Poisson and negative binomial models.
• Selector tables.
• One-way quartiles.
• Nonparametric Kruskal-Wallis test.
• Support for all comparisons for 2-by-M.
• Difference table and forest plot.
• A new Z Factor column in the Means Differences table.

Fit Response Screening functionality has been updated to include:

• Subgrouping.
• Setting Fixing Continuous Random Effects default to centered continuous.
• Selector tables.
• A new option for sliced effect comparisons within Multiple Comparisons when a model has interaction effects.
• A new option for sliced LS Means difference tests with random effects.
• New confidence intervals to LS Means and LS Means Differences.

Additional items

• Categorical platform enhancements include:

  • New summary statistics.
  • More format controls for the cross tab.
  • More false-discovery rate adjusted p-values.
  • Conditional highlighting.
  • Improved options for saving results in tables.
  • Enhancements for group comparisons.

• New Mean Difference table and forest plot in Categorical.

• Expanded comparison methods for Compare Slopes.

• K-Means clustering to specify an initial cluster means.

• Options in Model-Driven Multivariate Control Chart to add an out-of-control row indicator column, add a parallel plot with mean curve, detect a change point, and implement alarms/warnings.

• Model Screening enhancements: For K-Fold, the code now always creates three-way splits instead of two-way. For Nested K-Fold, the algorithm can now accommodate more cases, and defaults changed K to 5 (previously 4) and L to 4 (previously 5). Forest plot column for RASE to Summary results added.

Advanced statistical methods (JMP Pro)

Mixed Models

• The Stability Analysis option for Mixed Models enables you to estimate shelf life using mixed models. It includes an option to enable Extrapolation Control in the profiler for mixed models (SLS-REML, Mixed, and GLMM).
• Homogeneity of variance testing in Mixed Models.
• Support for Frequency and Weight Variable roles in both Mixed Models and GLMM.
• Sandwich estimator for covariance matrix in both Mixed Models and GLMM (for robust estimation of standard errors).
• Weight and Freq roles now in Mixed Model personality.
• Support for Containment DF in Fit Mixed and GLMM.

Structural Equation Modeling (SEM) enhancements

• New model shortcuts.
• New Model-Implied Instrumental Variables with Two-Stage Least Squares estimator.
• New option to generate R code from the SEM platform.
• Ability to test specific indirect effects.
• Improvements to survey reliability.
• Improvements to Assess Measurement Model and New Model-Implied Instrumental Variables Two-Stage Least Squares.

Life Sciences

• Distance Matrix platform: Compute a variety of distance matrices applicable to genomics and other wide data; perform related analysis such as principal components analysis.
• New Sequencing Variants add-in: Handle next generation sequencing files, extract genotype calls, and transform the data for analysis in JMP Pro.
• Marker imputation: Impute numeric missing marker genotypes.
• Marker admixture: Assess the probability of each sample having origin from a set of founding populations.
• Normalization: Prepare genomics data prior to statistical analysis. Normalization options include Row Standardize, CPM/RPM, CLR, Wrench, RPKM/FPKM, TPM, and TMM.
• Distance matrix: Compute a variety of distance matrices (Euclidian, Manhattan, Jaccard, Bray-Curtis, Gower, binary, and Hamming) applicable to genomics data, especially for microbiome.
• Offset variable role for GLMM.

Scripting, automation, and workflow

JMP Notebook

New scripting interface that provides a simple and easy to use way to create, document, organize, and test scripts in JMP using JSL, Python, or both.

JSL updates

Updates to JSL functionality include:

• Snippets of user-defined code can be quickly inserted into scripts.

• Box Model Padding and Chrome aligns the effect of JMP’s Display Box padding with other commonly used Box Models in other GUI toolkits.

• Column-Robust Report Customization creates a customized platform report that includes Send To Report and Dispatch calls in its script.

• Scripting experience enhancements include:

  • Script Editor preference page changes and new preferences.

  • Styling changes (namespaces/table names and scoped names).

  • Code folding improvements.

  • Toggle Code Folding menu item.

  • Python styling additions:

    • Brace matching.
    • Additional keyword coloring.
    • Attribute coloring.
    • Autocompletion while typing.
    • Highlighting instances of currently selected text.

• Workflow Builder can be searched by using keyboard shortcut Ctrl+F in name/notes/code editors to open a Find dialog that allows find/replace in the current editor or broadens the search to other steps in the workflow.

Additional items

• Add-in Builder automatically updates add-ins when adding and editing files.
• In-line text input in the JMP script editor is supported on Windows using the IME.
• Platform scriptable UI layout better delineates the boundaries of the sections of reports that are created by objects managed by JMP.
• Several tie break modes (average, minimum, maximum, row, arbitrary, and dense) have been added to Ranking and Ranking Tie functions.
• New JSL function supports multiple-response where clauses from data filter.
• New JSL functions Col Score, Col Interpolate, and Col N categories have been added.

Miscellaneous

JMP themes

Users now have the option to switch between three themes to customize the look and feel of JMP without making individual style adjustments.

Dark mode display setting

Option to use dark mode preference is now available on Windows.

Additional items

Other new features include:

• The option to cancel broadcast message.
• Better handling of color-blindness filters in color theme picker.
• The deprecation of San Francisco Crime and San Francisco Crime Distances data sets.
• Support for Save PDF preview.
• Spell checker.
• Exporting expression column images to an Excel spreadsheet.
• Manual Authorization window updated with information to help diagnose connection issues.

JMP Live

Additional JMP Live updates

• Linked data tables that allow dependent data tables to be set up so that they refresh when the larger table is updated by a data source.
• Ability to use pure Python in data refresh and import scripts.
• Support for multiple responses in Local Data Filters.
• Graphlets in hover labels that include them.
• Easier management of add-ins via the JMP Live web site.
• Custom report regeneration scripts.
• Support for profilers in the Bivariate, Distribution, Life Distribution, Fit Life by X, and Model Screening platforms.
• Additional options for reports in the Distribution, Bivariate, and One-way platforms.
• Support for 3D scatter plots.
• Ability to edit scripts that regenerate reports.
• Using paths to identify JMP Live objects in JSL instead of IDs.
• Resizing Graph Builder graphs and Control Chart reports to fit in a browser.
• Improved look for downloaded projects.
• Support for pie charts.
• Support for histograms containing multiple response data in the Distribution platform.
• Interactive word clouds in the Text Explorer platform.