turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Discussions
- :
- Discussions
- :
- Bagplot with jmp

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Oct 22, 2014 2:24 AM
(1139 views)

Hello,

i would like to know how to build a bagplot (see picture) using Jmp.

thank in advance for your help.

Best,

Medoune

4 REPLIES

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Oct 22, 2014 11:45 AM
(852 views)

Hello,

There is a JSL script available for what you want to do.

Have a look at the script and see if this does what you need.

Thanks,

Stan

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Oct 22, 2014 11:37 PM
(852 views)

Hello,

thank stanley.koprowski1 for your help. It is not exactly what i want to do. The plot i want have only two coordinates.

Thank

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Oct 23, 2014 1:17 AM
(852 views)

Hi medfand,

I'm sorry to say there isn't a built-in platform for bagplots in JMP. With a little bit of work you could script the polygons for a bagplot and layer them over a scatterplot -- if your comfortable with JSL that is pretty straightforward. But, I'm not sure what it would take to calculate the Tukey depth. In any case, bagplots are useful, and I'd be glad to pass along a suggestion to development on your behalf.

In the meantime, what is your main use of the bagplots? They're nice for a number of reasons, but perhaps there is already something in JMP that can help with the specific need you have. If you're interested in bivariate density, perhaps the non-parametic density contours would be useful?

Fit Y by X (bivariate) > RedTriange:Nonpar Density

That doesn't get you the outlier detection, but through another platform we could get both, and for more than two dimensions if we wish: Analyze:Multivariate Methods:Multivariate >> Red Triangle:Outlier Analysis:Mahalanobis Distances, and Scattplot Matrix Red Triangle:Nonpar Density.

If you haven't seen them before, the Mahalanobis distance is shown for each row in the dataset, and is a measure of multivariate distance to the multivariate mean formed by the k dimensions of Y specified, adjusting for the observed covariance structure. This is neat because it identifies points, not just for being far in distance, but those that don't make sense. For instance, petal width and height correlate, so it would be odd to observe a very wide but short pedal. This measure of distance picks up on that. In multivariate space adjusting for covariances that pedal is more extreme than a really wide and really tall pedal. In the screenshot below I selected those points above the upper control line.

Finally, you can script JMP to be a nice front-end for R and if you use bagplots a lot and want some pointers on building that let me know. It's easier than you might think. If you would like an example, here is a wrapper I wrote for a density overlay in GGPlot2 I like: JMP Wrapper for Overlaying Densities using R and GGplot2

I hope this helps some, even if not exactly what you were looking for!

Julian

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Oct 23, 2014 4:47 AM
(852 views)

Thank julian for your help. I already try to Fit Y by X (bivariate) > RedTriange:Nonpar Density but i d'ont have yet something like bagplot. Now i call R but it make time to run my script.

A bagplot is like a boxplot but in two dimensions. I really would like to do it with jmp because it allow me to win a lot of time.

Thanks, i really appreciate your help.

Best,