921424_en-US

liliconcom

Remove empty subplots in Group X graph

Good day,

I have a certain data of Y by X for a range of months and years. I would like to visualize this data without overlays, i.e. where each plot is located next to each other. (The reason is that real Y by X data may already contain several datasets overlayed).

Typically, I would add Year and Month as Group X variables. The problem is this creates a lot of blank "subplots" if there are no entries for each permutation of Group X variables. I would like to remove these empty "subplots".

 Other notes:

• I would like to use both Year and Month as Group X variables, and prefer not to create an additional variable "Year-Month".
• X and Y are typically continuous variables, Month and Year are typically nominal variables.

I'm using JMP 18.1.1.

Attached is the JMP table with a dummy data used for this example.

Data Exploration and VisualizationRe: Remove empty subplots in Group X graph

Creating combined column (like you suggested) is usually the easiest way to solve this fairly common issue you can (and will) come across with Graph Builder. There are some workarounds but they can be risky and will take some effort.

If I remember correctly, there are quite a few wishes about improving this in JMP Wish list.

Re: Remove empty subplots in Group X graph

A related wish, but not exactly the same:
 X group: restrict the values on the axis to the respective group 

Re: Remove empty subplots in Group X graph

For date values on the X axis, one can activate Label Row Nesting:

here is a wish to implement a similar functionality for Group X/Y:
Label Row Nesting for Group X 

Re: Remove empty subplots in Group X graph

Hello,

I've tested your problem.

If you only need to display data points and don't need the mean line, then follow hogi's suggestion and use nested variables on the X-axis; however, if you need the mean line, nesting doesn't seem to allow separating the mean line by month.

However, if you only need a nice graph, there is an alternative: use Measurement System Analysis (MSA) and make some simple settings. Hopefully, this will be helpful.

The method and results are as follows (Note: Since there are no repeated measurements, an error message indicating that the standard deviation cannot be calculated will appear; please ignore it):

Finally, it's been a long time since I've been to the community, and its language translation function seems to be even better now. I'm from China, and my English isn't very good. Previously, I needed a third-party translator to reply to posts and exchange ideas, but now it seems I can communicate across languages ​​directly without one. This is a very nice improvement.

Re: Remove empty subplots in Group X graph

Thank you for suggestions!

Indeed, if X variable is nominal, I can nest both group variables under X axis, which however connects points across the groups:

 

Using MSA platform to make a plot is an interesting idea, thanks!