As a SE, what is your JMP wish for the new year?
The What is your JMP wish for the new year? posting in the space is getting some good traffic and good suggestions.
I thought it would be good to give the SEs a place to post their wishes as well. Share your wishes in comments below.
If you see a wish you want too be sure to like it.
Happy holidays everyone!
SAMPLE DATA:
I have one more: Most of the sample data is described somewhere in the JMP books. But sometimes it is not easy to find this. So one sometimes get stucked with a data table and a saved script with a Report. No further Explanation or Highlights about the result. I imagine many users experience the same as I sometimes did: "ok this is a data table for that Topic and here's one script, and now? What's the clue?"
I'd like to have a script in each of the sample data tables referencing to the appropriate documentation of this specific example if available. If this is not available may we can include some short step notes to guide the People through the example. What is the base of the data, what is the Goal of the Analysis and what are the conclusions of the reports.
Re: As a SE, what is your JMP wish for the new year?Well, we already got a "Save Script to Data Table" Button in V12, so that whis is resolved,
My simple wish is for the addition of a "Cancel" button to the window this this message "JMP cannot replace the existing table because there are report windows that depend on the previous ordering."
I'd also like a kill button (short cut key) that would end all processes and do a memory dump.
Re: As a SE, what is your JMP wish for the new year?Dear Santa Claus,
I hope it is not too late for two more wishes (I hope it should be fine at least in East Coast time zone, where the wonderful JMP gifts are developed):
First, Structural Equation Modeling supported by JMP, with a GUI to specify and present models, would be great addition at least for many academic users. SEM is one of the remaining reasons to keep using SPSS tools.
Second, having a GUI language setting also in the Mac preferences would help a lot, especially in courses and workshops. It also would improve consistency between Mac and Win versions.
Now get ready for a very Happy Christmas,
Volker
Re: As a SE, what is your JMP wish for the new year?1. One of my favorite things to show new users of JMP is the ubiquitous selection -- select a point in a graph or table and it's selected everywhere. Simply amazing: the usually invisible relation between rows in the dataset and their representation in a graph manifests effortlessly, affording insights found nowhere else.
Deselecting points, however, can sometimes be frustrating when a large number of data points fill a particular visual (since there sometimes isn't a good place to "click off" of the selection, unless we return to the data table). As often as possible I prefer to use keyboard shortcuts to save time, and my wish is for a keyboard shortcut to clear the current row selection. I've made my own shortcut (with BetterTouchTool) by invoking "Select All" then "Inverse Selection," but it would be great to have a native shortcut like Shift-Command-A (or Shift-Control-A on the PC).
2. I'm often asked why there are no measures of effect sizes given for linear models in Fit Model or Fit Y by X. Most of these measures are trivial to compute by hand (assuming one knows which values to use), but I find myself reluctant to tell a potential customer they need to dust off their TI-83 to get Cohen's D.
3. "Set as default" for preferences in all platforms. This option is available in Categorical under the "Set Preferences" option, and it seems like a streamlined way to make JMP behave exactly the way one wants. I would love to launch Fit Y by X with a customer, go through turning on/off the output they wish, then select a "Make Default" option. Then, when we launch the platform with a new set of data they would see all their settings and customizations -- JMP exactly they way they like it.
4. "Embed HTML" option in a journal. After a workshop or webinar I like to share the journal I've been using, and an easy way to do that is exporting to HTML5. On occasion it would be helpful to embed a youtube video in the HTML5 output, but there is no easy method to embed that HTML in the journal before exporting to HTML5 (any HTML in a text box is displayed as text on the resulting webpage). Of course, a link to a video (or other website) can be included in the journal, but it would be snazzy to embed the video itself, to be revealed when someone opens the outline box containing the embed tag.
5. Recalling the specifications of an analysis is quick and easy in JMP (thank you, recall button and relaunch/redo analysis menu items!). In addition to these, I would find it useful if shift-clicking a platform would automatically recall and run with the last settings (without even showing the platform launch window). For instance, suppose I have Fit Y by X with two continuous variables. After looking quickly at the output I could close the window, make some exclusions in the dataset, and then I could then shift-click Fit Y by X in the menu and the bivariate output would be rerun without me having to click anything else. Yes, it only saves two button clicks each time (Recall & OK), but that adds up quickly.
6. Geocoding based on IP Address. One of the most used online survey tools, Qualtrics, automatically collects the IP address of survey respondents. If those IP addresses could be geocoded then anyone using qualtrics would immediately have the ability to geographically visualize survey responses.
7. Right-clicking any graph in any platform would give the option to "Open in Graph Builder" (assuming that graph type is supported). This would be useful when, say in Fit Model, you have a particular LS-Means plot you wish to further customize. One can easily save the LS-means table (right click -> Make into data table) then recreate the graphic in graph builder, but this would be yet another time saver, and useful for platform graphics that are not accompanied by a table.
I don't think I'm allowed any more wishes so I'll stop here JMP is the most amazing piece of software I've ever used and I'm thankful every day that it's my job to show people what it can do for them. If I could have just one wish (rather than my 7 above) it would be that everyone who could benefit from JMP would know about it. JMP doesn't just make statistical analysis easier, and it doesn't just make it faster, it makes it better. Much better. Given the central importance of data in today's world, anything that makes the analysis of data better is making the world a better place.
So, thanks everyone for making the world a better place.
Julian
Re: As a SE, what is your JMP wish for the new year?JMP is an amazing product, with many great new features in every release.
That said, my wish is to focus on existing features and functionality to improve stability, look and feel and most important consistency across platforms.
...and, the ability to easily set a random seed (wherever a random seed is used).
Thank you, and I hope everyone has a wonderful holiday season!
Mia
Re: As a SE, what is your JMP wish for the new year?My top wish is to make it as easy for non-native English speakers in Europe to import excel and txt files using European date formats, comma decimal seperators and other language differences in source files. These issues cause confusion with new non-English speaking users and can prevent them getting started with JMP. The impact of which is that they resort back to using Excel. See wish from Bernd and he can provide some example files if any of you would like to live a day in the life of a European non-English speaker.
Thanks, Merry Christmas and Happy New Year,
Malcolm
Re: As a SE, what is your JMP wish for the new year?Is it too late? In Stepwise, I'd like to see whole effects AND statistical heredity. And could you make it the default when there are categorical factors?
The antedependence and Toeplitz covariance structures for mixed models would be nice.
Support vector machines.
Re: As a SE, what is your JMP wish for the new year?I have one more, I have one more 
In Europe people have to work with data of all kind of formats. Decimal numbers, date- and time values are formatted in different ways. The only way to import data properly is to go to preferences and change the JMP language. Then you import the data and change the language back to the user interface that you are used to. We could save thousands of working hours if that language button was available in the import dialogues for Excel and CSV files.
I assumes that would boost European economy more than all the deficit spending activities of the European Central Bank.
Re: As a SE, what is your JMP wish for the new year?
Hello Jeff,
thank you for giviong us this opportunity to wish.
I'd like to see more options to visualize clinear constraints.
Linear constraints are saved as a tbale script either manually or by DOE. The profiler shows the dependencies and optimizes with resopect to the constraints. But none of the other profiler graphs limits factor combinations to the constraint area or displays the constraint either as a border line or a shaded area.
When I run a simulation experiment from a constraint optimization then the constraints are not carried over and I can vary the paramaters independently. Here as well factor ranges should be constrained. The space outside the constraints should be shaded in the contour profiler.
Re: As a SE, what is your JMP wish for the new year?My wish is world peace in 2015! In addition, I'd like to have Indicator Parameterization Estimates option for every Fit Model personality!
Currently this option is only available for Standard Least Squares personality. Except the Generalized Regression platform, the estimate for a nominal factor level reported by JMP has the interpretation as the differences in the predicted response for that level from the average predicted response over all levels.
Indicator Parameterization Estimates option such as the one in Standard Least Squares models enables the coefficient to be interpreted as the differences between the predicted response for that level and the predicted response for the last (i.e., suppressed) level. This is a more popular interpretation in regression with categorical factors, and the Generalized Regression does just that in the default Parameter Estimates for Original Predictors report.
Re: As a SE, what is your JMP wish for the new year?Thank you Jeff for soliciting our wishes!
Like Martin I have a couple of desires for added capabilities in JMP.
For one thing I would like to see JMP add more visual results to the Model Comparison platform. Model Comparison gives us a bar chart for R-Square, but what I’d also like to see is a “Visual Confusion Matrix” for Categorical Responses to go along with Misclassification rates – across all the models compared. Here’s what I currently make with the Scatterplot Matrix by Validation Group. This one compares predictions for six models of a binary response.
Scatterplot Matrix Crossvalidation=Test
For continuous responses I’d like to see the Actual vs. Prediction scatterplots with a line fit showing slope, R^2, and RMSE. I currently generate the following in Graph Builder to compare several models with a Column Switcher.
Graph Builder
(1)
The inclusion of a box in which to enter a “Power” under the scale button in the axis-settings dialog box was a great addition to JMP 11. It is so great to have the axes remained labeled in raw units while scaling the axis. I’d like to see the Power option also add a slider so that one can “swipe” through powers and watch the axis scale change and the spread of points change until the variance looks uniform – sort of a “Visual Box-Cox Transformation” tool. I imagine it would work like the stiffness slider for the Smoother in the Graph Builder. (At least one competitor has had this capability for 10 years.)
Thanks again for asking for our wishes.
Tom
Tom Donnelly, PhD, CAP
JMP Systems Engineer
& Co-insurrectionist
SAS Federal LLC
27 Farmingdale Lane • Newark, DE 19711 • USA
voice: 919-531-1380 • mobile: 302.489.9291 • fax: 919.677.4444
email: tom.donnelly@jmp.com
http://twitter.com/JMP_Fed
Re: As a SE, what is your JMP wish for the new year?I think this has been mentioned on the customer wish list as well but however:
Sometimes JMP freezes within a calculation or it just does not respond because of a (sometimes accidentially started time consuming Operation (e.g. with huge amount of data). It would really be great to Abort those calculations e.g. with Ctrl+C, to stop just that calcualtion and not whole JMP process. It happened more than once that some work got lost because of a never ending process and required restart of the whole JMP process.
A related Topic: Technical Support:
It would be great for customers to have a help menu item called Contact Support. This Feature should open up an email addressed to JMP Support (Support@jmp.com) with an attached file Version.txt gathering all JMP relevant Information as well potential log files or Crash files from that day and a possibility for the customer to describe the behaviour he wants to Report, e.g. with a predefined email text like:
1. What do you want to Report (Crash, unintended behaviour, enhancement, bug, ...)
PLEASE ENTER TEXT HERE
2. What's the behaviour in your own words
PLEASE ENTER TEXT HERE
3. What are the steps to reproduce this behaviour
PLEASE ENTER TEXT HERE
...
This would reduce the customer's reserve to contact us, we gather more Information about th euse and stability of JMP and for the sales and Service People we might get Information about customer issues much earlier, which probably help us to be proactive in our sales process. Asking if everything si ok after he submitted a Support request could be a good start to further improve the relationship and to decover pains and/or opportunities for new sales (after listening to the customer's pains )
Re: As a SE, what is your JMP wish for the new year?
- The DOE custom design Plattform is scriptable, the choice design platform not. That is, when I want to script a choice design (and not the choice modeling afterwards) this is currently not possible. As the custom design is not exactly tailorable to create a choice design it would be nice to have this, especially to e.g. also set the Prior mean and Prior variance. (I stumbled across this because of a customer I've been working with)
Re: As a SE, what is your JMP wish for the new year?
- In Journals when there are multiple e.g. text elements and I want to change the font type (size, italic, bold, whatever) I would like to select multiple elements and Change the font for all of them at once
- In reports I'm able to move parts of the Report to another place. This is great, but I'd like to have not just a copy of the Report which will added to another place but the Report itself. E.g. combining a Distribution Report with a bivariate fit next to each other. In the combined window I want a specific part of one Report, e.g. only the Distribution graph on another place, e.g. beyond the summary statistics. I can do this,but it is just a static graph and no more interactive. In Addition the original Distribution graph is still at the top and cannot be changed or deleted.
Re: As a SE, what is your JMP wish for the new year?
- I would like to have shortcuts for the JMP Debugger, like e.g. F10 for next step or F11 for step into and so forth. It is much more intuitive for programmers to use shortcuts rather than clicking.
