Log in to the Community to add your wish in a comment below.
"Like" the wishes from others that you want to see come true.
Hi all,
this is a wonderful idea, and I would like to post a "Graph Builder" wish related to the coloring of measures.
In the "Gradient Setting" I often have to change the default minimum + center + maximum setting to the 10% quantile + median + 90% quantile (I copy them from "distribution") to see more color-differentiation and better visualization of my data. This is not too bad - if I'm not working with local data filters with different quantiles ...
It would be just great to have the option to insert the quantile-percentages in alternative to the absolute values - or, even better, to switch entirely according to my "general settings".
All the best and a very good start in 2015,
Anja (anja.simon@dlh.de)
I wish that the Tabulate feature didn't change the "N" decimal places when I do a Change Format, use same decimal format, and set the rest of my table to one or two decimal places. Is there even an instance where I need to tell someone that my N is 149.00?
The ability to insert a column after the column currently selected would be nice.
Also, being able to search for variables in all windows in which variables are shown. It would also be nice to see long names in all variable windows as well.
It would be nice if JMP could connect the dots on an overlay plot in a sequence defined by numbers in a column. Presently if you have a hysteresis loop to plot you have to break the plot into sequential chunks and plot each of them to get a loop.
It would be most useful if there were a tool, similar to the one available for SAS users, that facilitated ingestion of XML data without having to write JSL.
Creating a network graph would be a wonderful addition to JMP.
Glad to hear that, Anja, because quantile-based gradient coloring is coming in JMP 12.
I'd like to see MDS (Multidimensional Scaling). There used to be a R addin for it, but now it's gone. Please bring it back!
Just great! But you have to leave something for Santa ….
Would love to have a native quantile regression fitting algorithm within JMP. There is one through a SAS connection but not very convenient. Should be easy to include, no?
Paul, loops and other order dependent shapes are possible in Overlay Plot if the Sort X option is not checked. (like the "Row Order" option in Graph Builder).
Run the code below with and without Sort X( 0 ) to see the difference:
New Table( "Tree",
Add Rows( 21 ),
New Column( "X",
Set Selected,
Set Values([10, 8, 8, 2, 6, 4, 8, 6, 10, 14, 12, 16, 14, 18, 12, 12, 10, 9, 10, 11, 8])),
New Column( "Y",
Set Values([0, 0, 5, 5, 8, 8, 13, 13, 18, 13, 13, 8, 8, 5, 5, 0, 0, ., ., ., .])),
New Column( "Y2",
Set Values([., ., ., ., ., ., 21, ., ., ., 20, ., ., ., ., ., ., 23, 18, 23, 21]))
)<<Overlay Plot(
X( :X ),
Y( :Y, :Y2 ),
Sort X( 0 ),
Connect Thru Missing( 1 ),
:Y( Connect Color( 52 ) ),
:Y2( Connect Color( 41 ) )
);
I have a related wish, though. An option for Graph Builder's Line graph, like in Overlay Plot, to "Connect Thru Missing" when Row Order is Checked. That would facilitate gap-less plotting of vertical profiles which typically in aquatic, atmospheric and soil sciences keep the y-axis reserved for depth/height and the x-axis for observed variables, which may be measured at different vertical resolution.
Your first wish is granted only you have to be sneaky and use the "add multiple columns", I know, you only want to add one but if you use the multiple option you can select the location of where you want to insert your one column.
Eh, kind of a non-solution if you have thousands of variables with fairly nondescript names. You then have to scroll your brains out trying to locate the exact location where you want the column to go. It's a painful process if you need to add in a bunch of columns in different locations.
That said, that is what I have been using since the base add column command in conjunction with the move column command is even more painful to deal with.
I want to wish the developers a wonderful holiday season and to thank them for taking the time to not only listen to our wishes but to work to grant our wishes!
Agreed the lots of columns are hard to deal with. I use column grouping to help manage such tables. However, If a column is selected and you used the "add multiple columns" it will automatically assume you want to insert your column(s) after the selected column.
Ah! I was not aware of that little tidbit. I will definitely try that out. Thank you!
You can actually already do that, by clicking Col-->Add multiple columns, and then select 1 column to insert, after the column currently selected. A bit counterintuitive, but it works
Edit: oops, didn't see but the question was already answered...
In the graph builder, I would like to be able to easily include sample size, and perhaps results of post hoc tests to ANOVA-style results (e.g. the connecting letters from Tukey test, above the boxplots or columns) - perhaps it is already possible, but I haven't found it yet...
I wish the Measurement Systems Analysis was more capable, flexible and the platform output was easier to comprehend. I get more complaints about this platform than any other; most people in my office will just take the variance components and copy them into Excel to complete the analysis there.
....I also wish there was a user friendly way to add spec limits by group for the variability charts. This can only be done now with some graphics scripting.
I'd love to see a history of my work presented in a mind map format. If I open a table, it shows up in a mind map. If I create a graph from that table, a link to that graph explaining the parameters used shows up. If I create another graph, a parallel link shows up. Sub table, another parallel link. Then links from that sub table to anything built from it.
I write scripts that are executed both in Europe AND the US of A. The problem is that the decimal sign is different, which means I have to add an input parameter defining the decimal sign. To solve this I have to spent hours of writing code to get around this.
P L E A S E, make it so that there is an expression to set the decimal sign default on a point or comma, from which on the script uses a point or a comma from that point on (until defined otherwise). Something like
Set Decimal Point ( 1 );
Thank you!
Jan
There are times when I do NOT want a new data table, like when I sort the data using a certain column. There should be an option as to whether you want a new table or not.
Would like to build control charts with seasonal data - see Wheeler's published dicussion on how to do this.
Two wishes:
1. Would like to have a variable viewer in the script editor similar to what is in the debugger.
2. A "graph builder" for spectroscopic or x-ray diffraction type data that would allow fitting of multiple gaussian, lorentzian, etc... peaks.
Thanks!
A good addition would be the ability to highlight all instances of a variable in a script when you double click on one instance. Similar to the feature Notepad++. When your scripts delve into the thousandth line it gets pretty cumbersome trying to find the inital initiation of a specific variabe. Yes, I know Ctrl+F exists but being able to scroll to see all instances can be pretty helpful while debugging!
A zoom for table views would enable to get an overview of data quality. In a large table there could be hidden that too many values are missing for a reliable statistic.
Today I export often to EXCEL and tuen zoom ti 10% or 25% to see the table as a whole and find the gaps.
Then I can decide if the missing data jeopardizes reliable results.
Have you utilized the Tables>Missing Data Pattern and run the Treemap script?
You won’t get a new datatable if there are no platforms created with the original one. Do your sorting first, then build the platform.
My #1 wish is that JMP stability improve. It crashes for me sometimes when I resize windows, select items in a platform, just randomly... I've never used a professionally written program that crashes anywhere near as often as this one. Very frustrating, and responsible for hours of lost work probably every week.
