JMP Wish List

I would like to have possibility to use similar visualization as PowerBI provides with Ribbon Chart which is quite close to ordered stacked bar chart with some visualization between different bars (easier to see changes over x-axis than cluster bar charts). Some links regarding the PowerBI version Create ribbon charts in Power BI (learn.microsoft.com) and Ribbon Chart is the Next Generation of Stacked Column Chart (radacad.com) .   Below is an example using JMP's Failures sample data: JMP stacked bar chart: JMP with lines:   JMP stacked bar chart with lines: JMP stacked area chart: JMP data split and parallel plot with columns set as Ordinal: JMP split data with lines:   PowerBI stacked bar chart: PowerBI ribbon chart with connectors set invisible (ordered stacked bar chart):   PowerBI ribbon chart:   Out of these I prefer Ribbon chart as it clearly shows if ordering has changed in x-axis (it also has quite useful tooltips which shows the change).  I tried to find what name that chart usually has and I came across Bump Chart, area bump chart, alluvial diagram and sankey bump chart. Bump chart and alluvial seem to be quite different but I think they are also visualizations that could have their place JMP's Graph Builder. ... View more

☐ cool new feature ☐ could help many users! ☐ removes a „bug“ ☑ nice to have ☐ nobody needs it     Enabling the option "conditional" makes a data filter extremely user friendly. By consecutive selection of items in the filter lists, the user defines a hierarchical order how filters react on each other. The selection in the fist filter determines the subset that is displayed in the next filter and so on ...   There are some limitations which could be fixed in a future release of jmp: - conditional data filters are not possible for columns with propery multiple response: https://community.jmp.com/t5/Discussions/Data-Filter-you-have-to-choose-between-quot-conditional-quot-amp/m-p/555435/highlight/true#M76974  - same for text filters: https://community.jmp.com/t5/Discussions/Data-Filter-you-have-to-choose-between-quot-conditional-quot-amp/m-p/556116/highlight/true#M77008 - with the option "conditional" enabled, multiple data filter groups are not possible  https://community.jmp.com/t5/Discussions/Data-Filter-you-have-to-choose-between-quot-conditional-quot-amp/m-p/555435/highlight/true#M76974    more wishes submitted by   ... View more

I am currently using JMP 16 and use the graph builder to group data for reporting and presentations.  When the data is subgrouped in the Group Y or Group X axes, the font size and type is locked.  This makes it difficult to read when reporting.   Thank you, JasonD ... View more

Hi.  I propose there should be an optional argument that forces all columns to be Character Data Type for both the Open() and Multiple File Import() commands.  This request is similar to, but not exactly the same, as these 2 requests:   https://community.jmp.com/t5/JMP-Wish-List/Excel-Import-specific-column-types-or-formats/idi-p/248136  https://community.jmp.com/t5/Discussions/Force-column-from-CSV-import-as-Character-not-Numeric-Date/m-p/260752#M51078    The first link requests the ability to manually specify column data type for Excel files in the Open() GUI, whereas my request is addressing usage of the Open() command in a script, not the GUI, and applies to both csv and xlsx files.  The second link appears to apply only to MFI() and csv files, whereas, I would like this ability for both MFI() and Open(), and for both csv and xlsx files.   I realize there exist various scripting methods that could do this for certain scenarios, although not all. Some of these methods might require you to know the column names/locations in advance, or require that the column names always be on the first row.  This is not something we necessarily have control over.     I feel having a single option to set would simplify this greatly for many users.  Also, this would save time since JMP would not need to spend time scanning the file to detect column data type.  Lastly, my feeling is that this could be low-hanging fruit to implement, in that you'd just be turning off the portion of the internal code that scans each column.   Hope this is possible.  Thanks. ... View more

What inspired this wish list request?   In JMP version 17, it is quite time consuming to randomly fit numerous neural net models to hopefully find one that fits well.  If we don't type in the seed number at the start we cannot reproduce the exact analysis later, so we cannot currently reproduce the fit versus validation plots and prediction profilers.     What is the improvement you would like to see?            Why is this idea important? Development of neural net models in JMP is a potentially powerful feature, but it needs improvement to make the modeler's job a bit less repetitive in searching for a neural net that well fits the data.       ... View more

If development of Script Editor will continue (and I hope it will, or hopefully JMP/SAS will create functionality/support to some other code editor) add support for code snippets. These make writing code much faster and will reduce the amount of typos. One implementation example of code snippets can be seen for example here Snippets in Visual Studio Code  ... View more

It would be quite convenient if JMP allowed the Substitute() function to have an optional argument such as Substitute( ...., First Occurrence( 1 ) ) .  This will help when passing arguments into a function by reference.  Consider the following: some function = Function( {_function, _args}, {Default Local}, function = Name Expr( _function ); Eval( Substitute( _args, {}, Expr( function ) ) ) ); other function = Function( {_1, _2, _3 = "optional"}, Show( _1, _2, _3 ) ); some function( Name Expr( other function ), {1, 2, 3} ); /* OUTPUT _1 = 1; _2 = 2; _3 = 3; */ some function( Name Expr( other function ), {1, {2, 4}, 3} ); /* EXPECTED OUTPUT _1 = 1; _2 = {2, 4}; _3 = 3; */ Here in the first function I am calling a referenced function with a given argument list.  The problem is that the Subsitute() replaces all lists with the function call, instead of the outer one only.  This is where it would be nice to substitute only the first occurrence of the List() expression. ... View more

A process map or value stream map builder feature would be awesome.  The user would enter the process step name and have the option to enter additional elements like factors, responses, noise, specifications, levels, cycle times, etc. for each step.  The user could also select factors and create a DOE or link process steps to capability and control chart data sets to identify bottlenecks, critical chains, or areas for improvement.  JMP could also develop analysis platforms focused on the relationships between process steps.  ... View more

I'd like to suggest the following for equivalence testing and practical difference testing: Add a column property for practical difference limits.  This will be reference whenever an equivalence test is done in Distribution, Fit Y by X, and Response Screener In Distribution, Fit Y by X, and Fit Model, add in a report out option for practical significance.  This would be in line with the report out options you can get from Response Screener -> save compare means. I'd also like to see equivalency testing and practical difference testing for Stdev's (variances would be fine as well)   ... View more

In an unequal slopes ANCOVA model the individual slopes can be estimated with the Estimates > Custom test or Estimates > Compare slopes option in the output for Fit model. Standard errors and p-values are also given but I miss confidence intervals for the slopes. This should also be possible if additional random effects are added to the model (a mixed model). ... View more

Hello,   I'm currently using the EMP platform from jmp to determine ICC's for: 30 repeated measures, 1 rater, 10 participants. Is is possible to deduct CI's for the ICC's?   Thank you,   Sarah ... View more

What inspired this wish list request? Please describe the current issue that needs improvement or the problem to be solved that is not easy or possible right now, with an example use case.    Hello, I am trying to put together some complex charts for my company and I have run into an issue. I have to show a certain value for multiple categories over the span of a dozen quarters (see image below). I want to be able to order the categories in my stacked bar chart by the value in my y-axis. In this example, I would want the non-fail category to be on the bottom of the stack followed by application category and so on and so forth.   I know that I can manually assign the stack order but that is not robust enough for my purposes. I have local data filters that, when used, will change the value of the categories and then they will need to be manually restacked again.   I have talked with the JMP experts assigned to my company and they told me that there wasn't a built in functionality for this yet which is why I am writing this post.       What is the improvement you would like to see? Please describe the idea for improving JMP. Please include mock-ups, wireframes, screenshots, scripts, other documents or examples from other software that help describe the change you would like to see.    Just like how you are able to order the x axis bars of the bar chart by descending/ascending order of the overlay values (see below). I want to be able to order the overlay stack in the same manner.       Why is this idea important? Please describe the value to you and/or other users if the idea is implemented (for example, ease of use, must have,…).   This seems like a must have for a graph builder. Being able to order your data by values that are displayed on the graph should be easy to do and is something I would think a lot of people would expect JMP to have. ... View more

What inspired this wish list request? Messages sent to the log can be slow and also confusing to users   What is the improvement you would like to see? Don't swamp the log with useless messages   Why is this idea important? It's more of a nuisance, really.   Consider the following JSL -- I am attaching a function to the name "print" within a custom namespace. New Namespace( "This is a NAMESPACE -- it is different than BUILTIN!!!!!!!!!!!!!!!" ); "This is a NAMESPACE -- it is different than BUILTIN!!!!!!!!!!!!!!!":print = Function( {}, Print( "Why does JMP think I am trying to overwrite a Builtin???" ) )   Why do I get this warning?  I am not trying to overwrite a Builtin.  I know that I am putting the name within a namespace and that is the way that I will be accessing the value held by that name.   Can the JMP compiler be aware of when the name is within a namespace and not send these messages to the log?  I use namespaces to create custom classes, so whenever a new instance is created it will spam the log with these superfluous messages.     ... View more

What inspired this wish list request?  Process characterization reports often include prediction formulas and simulations.      What is the improvement you would like to see? Option to include prediction formulas in the easy DOE report export. Also would like to be able to export models/prediction and PredSE formulas to a jmp table from the easyDOE.   Why is this idea important? Because models made by an easyDOE user may need more exploration by advanced users. For example, comparing an easyDOE model to one made a different way, simulations, simple plots against data, etc. Typing prediction formulas with lots of factors is other non-trivial.      ... View more

What inspired this wish list request?  My companies' process control groups are increasingly using the Process Screening platform and the alarm rate is becoming a key metric for us.  Also, we are using three-way chart more and more, and it is just as important on three-way charts to track and understand the alarm rate on the within charts as the mean charts.  When trying to compare results across platforms as well as expected results, we found discrepancies.   I've been working with @PatrickGiuliano at JMP Technical Support to work through how the platform has been designed to function vs this requests goals.   What is the improvement you would like to see?  The alarm rate section of the Process Screening (PS) platform is a great way to be able to pareto out what process to attack first as well as track performance changes over time while also being able to quickly visualize trends.  I would like to see the following: The "Alarm Rate Column" and "Any Alarm" columns should respect all selected alarm rates as calculated off both upper and lower control limits.  Currently they only calculate off both control limits for the mean chart (as well as any other tests) but only the upper control limit for the "Range Limit Exceeded" when a three-way chart is selected. Not respecting the LCL causes different alarm rates when viewing a process in PS vs Control Chart builder.  This also creates control charts that are different looking than Control Chart builder when using "show chart as selected" in the PS and adding in the dispersion chart as only the UCL is shown. The names for the dispersion charts should be changed to be clearer.  "Range Limit Exceeded" should go to "Disp Limit Exceeded" and be used with MR/-R/-S/three way S/R charts and "3W MR Limit exceeded" should be used for MR charts on three-way charts. The alarm rates and alarm count should not be the sum of the alarm rates for each "Test" column and three way "Range Limit Exceeded".  They should only sum up across a subgroup (so if any subgroup fails a "test" on the mean chart or the subgroup dispersion chart exceeds a Control limit, that should only +1 the value on the "Any Alarm" column and only add +1 to the numerator in the "Alarm rate" column.  For example, if subgroup 2 (of 10) fails for the mean and 3 way within dispersion chart UCL's, and is the only subgroup to fail then the "Alarm Rate" should be 10% and the "Any Alarm" should be 1.  Currently this would show as 20% and 2, which is a misleading look at the process. Since the goal of Process Screening is to be a kind of one-stop shop for an overview of all your processes, it would be value-added to the users to have the platform have the overall alarm rate reflect all the alarm types being selected by the user.  (So, I would like to see a Means alarm rate, MR alarm rate, Range Within alarm rate, and an Overall alarm rate which flags for any tests that were selected by the users to flag a subgroup as defective {MR for an I-MR chart can be excluded here as it spans "two subgroups}).                 How this might look:     Note: took keep consistent with Control Chart Builder using “Test All Beyond Limits - All” vs "Enable all tests" might work better.  I noticed that after I made this picture above.   In the picture below you can see both issues for items 1 and 3: There are no lower control limits on either of the dispersion charts and the platform does not respect any violations of the LCL (which is an issue for -R or -S or the within charts in three-way charts) and thus, leads to a incorrect alarm rate. For the electricity subgroup, it is out of control for both the X-Bar and Range within charts and thus is counted as two alarms, when it should only be counted as one (as the denominator is the number of subgroups).  Thus, the alarm rate should be 3/17 not 4/17.   If for some reason item 1 from above cannot be added on, at the very least, then the names should be changed to “Upper moving range limit exceeded” and “Upper range limit exceeded” so that the user won't assume that they include both due to the current naming convention.       Why is this idea important?  This is important for the following reasons: You should get the same results from both Control Chart Builder and Process Screening, especially since Process Screening will send you to Control Chart builder as an available feature in the platform. Tracking and understanding within process variation is just as important as variation between the means and is especially critical when using three-way charts. Different companies will have different priorities on what alarm rates matter and thus the user group should be able to control what to display in the Process Screening platform (e.g. whether to turn on or off LCLs in MR and R charts). Whew, hopefully that came out clear.  Please let me know if I need to clear anything up!   Steve   ... View more

What inspired this wish list request?  When making a new data table I frequently use code dt = new data table() but this throws an error since only new table() works.  Tables are frequently called data tables in documentation and other functions and would be good to support new data table() as an alias for new table().   What is the improvement you would like to see?  Please make a function "New Data Table()" as alias for "New Table()"   Why is this idea important?  Would remove a frequent error when coding in JSL.   I am using JMP 16.2 ... View more

Currently if you want to add column to a function used in Formula Editor you have to either Drag and drop the column to correct place Select the place you want to add the column into and then double click the column to add it Open the formula as JSL script and write the column reference My suggestion is that new option should be added: right click selection for the columns (in best case only valid columns would be showed in the list). Right click menu would make it easier and faster to add new columns to functions, especially when you are writing more complicated ones / want to edit formulas.   JMP does already offer quite a few options in the right-click menu for Formula Editor, but I think column selection would fit there within the first "block". Quick paint example: ... View more

I'd like to suggest that the behavior of the control boxes for the elements active in graph builder change so that if you have minimized any of the boxes, when a new element is added they stay minimized.  Right now if you spend the time minimizing the boxes as soon as you add a new element, they all expand.   Two other suggestions would be for efficiency would be: 1. add the ability to hold down shift and ctrl at the same time and collapse all the control boxes (right now the best you can do is hold down control collapse/expand boxes of the same type) 2. add the ability to hold down control and broadcast the setting change of a control box to all other control boxes of that type.   Thanks for considering these requests.   Steve ... View more

I'd like to see a JSL code option that would return the coordinates of the mouse so that a new window can be position in reference to the mouse location.  This will allow the new window to be located close to the mouse for quick access of options on the window.  How I will be using this is  a new window will be generated that contains scripting buttons for often used actions and thus having it appear near the mouse will allow an efficiency gain vs having to move the mouse up to the toolbar (the bigger the screen the more this can matter!).   Thanks for considering this request.   Steve ... View more