JMP® Marketing Features
1) Query Builder. A modern SQL database query tool helps users discover interesting data and make correct queries that are repeatable and sharable.
Product: JMP
Audience: Scientists and engineers who need to access data in databases.
Business Pains Addressed:
- Large amount of time spent in data preparation.
- Too much IT staff time supporting query needs of analysts.
- Inability to join tables without having to write custom SQL code. Automatically finds the outer join keys (for selected tables).
- Having to remember data values for filtering rather than being able to pick from a list.
- Having to build custom user interfaces to prompt the user for filter values and substitute the values in SQL expressions.
- Address competition with Spotfire and Tableau.
Key Customer: Micron, Intel, P&G, Cree, Freescale, Abbott, Intersil
2) Interactive model editing. Lets you quickly run and compare various models with a drag-and-drop interface within the report. Speeds up the analysis, getting you to the model that performs well in fewer clicks, and emphasizes the interactive nature of statistical analysis with JMP. Supported in many Fit Model personalities.
Product: JMP
Audience: People performing exploratory modeling.
Business Pains Addressed:
- After fitting a model, you need to add or remove terms, drop terms from a model that are not significant or interesting, or make the model more parsimonious. Previously, this required you to enter the Fit Model launch dialog. Now you can do it in the analysis flow, without leaving the platform.
- P-values are often too small to format, and larger values are preferred in graphs to mean more significant. LogWorths are thus used to display effect importance in a bar chart to aid with effect selection.
Key Customer: P&G.
3) PPT Export. It is always nice to show your data analysis results from JMP directly in JMP. But in some cases, you may need to prepare a slide deck to present your information to others. Now, JMP makes it easy to export your reports as PowerPoint slides. You can reformat slide decks created by JMP with a company template. Tables include editable text, and JMP automatically transfers graph titles to slide titles.
Product: JMP
Audience: All JMP users who need to present results in PowerPoint.
Business Pains Addressed:
- Getting high-quality graphical output from JMP to PowerPoint is a difficult manual process.
- You must manually type graph titles into slide titles, potentially causing a double title on a slide, wasting space in the process.
- Tables either come in as graphics, which you cannot edit, or their formatting may be wrong.
- Competitors, such as Minitab, make exporting results to PowerPoint easy, often in a single click – now JMP provides the same one-click export ability to PowerPoint.
Key Customers: CAB, First Solar, JMP Clinical, Freescale, Abbott, Bill Meeker, GSK, P&G.
4) Images in Data Tables. Expression data type added to data table with good support for images. Create columns of images directly from Graph Builder. JMP is the only fully-featured statistical analysis tool that supports this functionality.
Product: JMP
Audience: Anyone who wishes to add images to data tables, or include expressions in a column.
Business Pains Addressed:
- Horizontal resizing of data tables not supported.
- Data tables did not support images or expressions well.
5) Recode. Data cleaning/preparation prior to modeling can take up to three times as long as the modeling or analysis itself. Recoding columns that contain character data is often responsible for much of the preparation time. Categorical data is often messy due to misspellings, inconsistent capitalization, abbreviations and structure that require cleaning in preparation for successful modeling.
Product: JMP
Audience: Anyone preparing data for analysis.
Business Pains Addressed:
- Recoding categorical data has tended to be a tedious and manual process. Now, JMP lets you clean up messy data with a suite of powerful tools, including string clustering based on character edit distance. You can recode manually or use an automatic recode algorithm.
- Pains around manual grouping, filtering, searching and sorts columns with many levels of character data.
6) Graph Builder Improvements. Many improvements to Graph Builder, including: supporting selections in legend; more control of gaps between graph panels; graph matrix; right-clicking gap now supports color, transparency and size; and the graph borders. Improvements also include: smoothly connected lines; overlaid histograms; new bar element options for needle and float; support for step, horizontal and vertical modes in line element; and pages role (a pseudo-BY variable); better support for coloring graphs by various color gradients. Finally, a Graph Builder launch dialog is available through script > Relaunch Analysis or with JSL through Graph Builder (Launch Dialog).
Product: JMP
Audience: Users performing exploratory data analysis or data visualization.
Business Pains Addressed:
- Some JMP users still use the Chart platform because there are features that Graph Builder does not support.
- Sometimes it is desirable to generate a graph in Graph Builder using a launch dialog, and not with drag and drop (for large data problems, or when a graph matrix is desired).
7) Interactive HTML output for Profiler and Bubble Plot. Profiler and Bubble Plot are two of the most popular tools for communicating results with others outside of JMP. Previously, interactive versions of these platforms were available only as Flash, which many devices (especially mobile) cannot play.
Product: JMP
Audience: Anyone sharing or communicating results who may not have or have access to JMP.
Business Pains Addressed:
- Only static versions of Profiler and Bubble Plot were available previously, limiting their usability as a communication tool outside of JMP.
- Flash versions of these two reports are not well-supported and prone to deployment difficulties, especially on mobile devices.
8)Selection Filtering. Row selection performed in one platform is used to filter the view in another platform to include only the selected rows.
Product: JMP
Intended Audience: Customers building dashboard-style reports.
Business Pains Addressed: Customers often have in-the-flow questions that are spurred by an analysis, and they need to further analyze a subset of their data. The Selection Filter makes it easy to incorporate user-defined analysis flows into a dashboard. Solutions that once required scripting with subset tables can now be completed by dragging and dropping.
Key customers: This has come up in CAB meetings in the past, and Jim Nelson at Freescale shared a JSL example that he had developed.
9) General DOE Enhancements.
- Blocking in DSDs
- Enhancements to Space-Filling Designs – Qualitative Factors, Augment, MaxPro – new criterion for building these designs.
- Split Plots with covariate factors.
- A new wizard in the Custom Design platform that allows you use a Data Filter-like interface for chosen specific factor combinations that should not be considered in a design.
Product: JMP
Audience: Anyone performing DOEs.
Business Pains Addressed:
- Previously disallowed combinations required writing an expression and knowing Boolean logic. This was challenging to many users.
- Blocking was not supported when creating DSDs.
- Split plots did not support covariate factors.
- Space-Filling Designs could not be augmented.
10) Multiple Correspondence Analysis. Multiple Correspondence Analysis is a data analysis technique that’s similar to principal component analysis, but for categorical data. It helps to visualize and uncover hidden structures in a data set consisting of more than two categorical columns. Such information is useful in market research to identify target audiences for advertisements.
Product: JMP
Audience: Consumer/market researchers, behavioral scientists.
Business Pains Addressed:
- It was sometimes difficult to identify why multiple individuals answered questions in a similar way or to identify groups of individuals based on their characteristics.
- It was difficult to visualize the associations between multiple categorical variables or multiple answers data.
Key Customers: France, Japan
11) Process Capability. A unified platform for assessing process capability. JMP reports short-term capability as “within capability” and long-term capability as “overall performance.” You can subgroup a process by nesting the column in a categorical variable, or set a constant subgroup size for a process. The graphics are updated, and you see both CPK and PPK by default.
Product: JMP
Audience: Quality engineers and Six Sigma practitioners.
Business Pains Addressed:
- Process capability now is its own platform in the quality and process menu, where you would expect to find it.
- Streamlined creation of spec limits.
Key Customers: Freescale, Amgen, Gore
12) Destructive Degradation. To measure a product characteristic, sometimes the product must be destroyed. For example, when measuring breakting strength, the product is stressed until it breaks. The new Destructive Degradation platform for reliability studies handles this type of analysis and has a built-in model library with many standard models, transformations and data distribution. Includes a model comparison, CDF and Quantile Profilers.
Product: JMP
Audience: Reliability Engineers.
Business Pains Addressed:
- Destructive Degradation was hard to find in previous versions of JMP, but now it is its own platform.
- Streamlined launch now lets users pick degradation paths from a list.
Key Customers: P&G, Bill Meeker
13) Clustering Improvements. An updated Cluster platform helps address common pains in data preparation prior to clustering. This includes reporting, data manipulation, reducing data dimensionality and data summarization.
Product: JMP
Audience: Semiconductor industry, engineers.
Business Pains Addressed:
- Reporting: To characterize clusters, you had to save the cluster ID, then prepare a summary, and then make a graph. This happens in one step now.
- Data manipulation: Data often needs manipulation prior to clustering. The new platform automatically transposes, stacks or splits data, letting you process raw data in-line with the analysis and saving many steps in the process.
- Dimensionality: In the semiconductor industry and others, it is important to determine clusters of wafer defect patterns. But this previously meant creating one variable to make XY coordinates, then splitting on this dimension, then clustering and then merging the clusters. Now, JMP accomplishes this in one step.
- Summarization: Often you want to cluster summarized data, and it takes multiple steps.
- Scoring: You sometimes need a scoring formula to classify new data into clusters.
Key Customers: Micron
14) Discriminant Improvements. The updated Discriminant platform includes a number of new features, making it a useful tool in a predictive modeler’s toolbox. New features include:
- Cross-validation support. (JMP Pro Only)
- Linear Wide, a new method for wide data cases. Useful when the covariance matrix is too big to calculate (1000s+ variables).
- Make Scoring Script lets you create a script that has the same effect as the Save command if you run it, but a script can be applied to a separate data table.
Product: JMP
Audience: Semiconductor industry, engineers, predictive modelers, scientists.
Business Pains Addressed:
- Discriminant analysis should be considered a viable tool in predictive modeling for dealing with big-data predictive models and also for better handling wide data cases.
- New capabilities in the Discriminant platform enhance its usefulness for predictive modeling and big data problems.
Key Customers: Micron
15) Modeling Utilities. A number of utilities help prepare data sets for analysis automatically, saving time in data preparation, where you spend most data analysis time.
Product: JMP
Audience: Anyone preparing data for modeling.
Business Pains Addressed:
- Data often contains missing values that must be imputed. Now, a unified missing value utility in JMP lets you impute missing data using a number of methods; count or cluster missing values; and perform row selection, coloring, exclusion or cell coloring. By streamlining data pre-processing, JMP saves valuable analysis time.(Explore Missing Values)
- When you have hundreds of columns as possible predictors and you want to try them all, you need some way to cull the bad ones or elevate the good ones, speeding the subsequent modeling tasks in the process. (Screen Predictors)
- Data sets are often full of outliers, which you must deal with prior to modeling. This can be a manual, tedious and time-consuming task. Now, in one click, you can identify things like outliers and error codes, and then deal with them quickly using one of a number of techniques. (Explore Outliers)
JMP® Pro Marketing Features
1) Factor-Covering Arrays. Covering arrays are used in testing applications where factor interactions may lead to failures. You may want to design an experiment to maximize the probability of finding defects while also minimizing cost and time. Covering arrays let you do just that. And when there are combinations of factors that create implausible conditions, you can use the new and interactive Disallowed Combinations wizard to automatically exclude these combinations of factor settings from the design.
Product: JMP Pro
Audience: Software testers and developers.
Business Pains Addressed:
- To design an experiment for factor covering, you previously needed to use another software package; now you can design these experiments in JMP.
- With some problems, other software packages do not provide the global minimum number of runs, potentially increasing the cost of running the experiment.
- Disallowed combinations can be difficult to write algebraically, especially in these types of experiments, which may have many categorical factors with many levels.
2) Enhancements to Advanced Analytics Modeling. Generalized Regression, a Fit Model personality, has a new graphical/interaction model selection and informative diagnostics bundle. And it now boasts up to 1,000-times better performance. Additionally, you can modify tuning parameters and select new distributions, such as Cauchy, Beta, Exponential, Zero-Inflated Gamma, Beta Binomial and Zero-Inflated Beta Binomial. The feature now includes support for effect heredity and inverse prediction and forced effects. Mixed Model, a Fit Model personality, has a new optimizer for greatly improved performance on larger problems.
Product: JMP Pro
Audience: Predictive modelers, users of advanced analytics, engineers and scientists dealing with messy or large data problems.
Business Pains Addressed:
- Fit Generalized, Fit Mixed and Robust Methods were often slow to fit larger problems or complex problems with higher-order effects.
- Fit Generalized lacked diagnostic plots and reports that other Fit Model personalities provided.
- Fit Generalized lacked an interactive way to adjust the model fit.
- Quantile regression was previously available as a SAS Add-In, which was removed in JMP 11.1. Now we can perform this statistical analysis natively in JMP Pro.
3) PLS-DA and assigning measures of precision to model predictions in PLS. PLS discriminant analysis (PLS-DA) fits a PLS regression with a categorical response. The advantage of this classification method compared to GLM or DA is that it handles correlated factors, missing data and can model more factors than observations. And there are almost no assumptions that must be made about the data prior to modeling. Additionally, bootstrapping the model predictions is now supported for all PLS model fits.
Product: JMP Pro
Audience: Modelers deal with wide problems; more columns than observations
Business Pains Addressed:
- Customers need to fit PLS models to categorical responses. Often have more factors than observations (wide and short data tables) where other modeling won’t work.
- Often these data tables include correlated factors, which cause other modeling methods to fail.
- JMP Pro 12 introduces bootstrapping for PLS model results.
Key Customers: Syngenta, Unilever, Ian Cox, P&G
4)Create Validation Column Modeling Utility. Creates validation columns with a single click either using pure random technique, or stratified to ensure that all categories in the stratified column are included in the training set.
Product: JMP Pro
Audience: Predictive modelers, anyone looking for an easy way to create a validation column for honest assessment of model performance.
Business Pains Addressed:
- Validation column creation was largely a manual process that required multiple clicks to divide a data set into train, validation and test sets.
If you required more than a simple random sample of the data for splitting, you had to use an add-in or other techniques to generate the best splits. Now you can split data based on the problem, using one of several included algorithms.
