cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
  • Image to Table JMP extension imports single or batch images as data tables. Useful to extract curves from pictures or convert image data to curve data to use in analyses like Functional Data Explorer. Download and install the extension.

Discussions

Solve problems, and share tips and tricks with other JMP users.
Choose Language Hide Translation Bar
renvik04
Level I

How Do You Handle Missing Values Without Losing Important Insights in Your Data?

I'm interested in learning how other JMP users deal with missing values during data analysis.

Removing incomplete observations might seem like the easiest solution, but it can sometimes affect the accuracy of results or hide useful patterns.

When working with datasets that contain missing values, do you prefer imputation, excluding certain observations, or investigating why the data is missing in the first place?

I'm particularly interested in how you decide which approach is appropriate for different types of analysis.

For those using JMP, are there specific tools or features that make this process easier?

I'd appreciate hearing about your practical experiences and recommendations.

3 REPLIES 3
statman
Super User

Re: How Do You Handle Missing Values Without Losing Important Insights in Your Data?

Much depends on how the data was acquired. Is this observational data or data that comes from directed sampling or DOE?

In any case, you should be curious as to why the data is missing. I have specific options I use for data that I planned to acquire.

"All models are wrong, some are useful" G.E.P. Box
P_Bartell
Level VIII

Re: How Do You Handle Missing Values Without Losing Important Insights in Your Data?

To add a few thoughts to what @statman has contributed, there are lots of ways to handle the analytics piece of the issue of missing data, many you suggest, and IMO each has their appropriate place under the sun...but IMO the 'best' pathway is very contextually and practically dependent so it's hard to generalize or make specific recommendations. For example, you run a DOE and have missing values. Does the extent of the 'missingness' make achieving your experimental goals and objectives problematic? If so, I'd bite the bullet and go back and get that data (as opposed to imputation, leave it out, etc.) . If the 'missingness is no big deal wrt to your experimental goals and objectives, well just leave it out or impute...try both and see what happens?

Another place to start is investigating the 'informative missing' capabilities in JMP described here in the JMP Documentation: Informative Missing in JMP . An example is shown here: Example of Informative Missing . 

prowheysa
New Member

Re: How Do You Handle Missing Values Without Losing Important Insights in Your Data?

good

Recommended Articles