Jump to content

Trellis ANOVA


TannerE

Recommended Posts

There's probably a way to do this. I just don't know what else to call it other than to trellis the anova. Long story short, I have three essential columns : appliance, test, and value. Generally speaking, appliance would be my x value for the data relationships Anova analysis and value would be the y-axis. The difficulty is that each test has completely different groupings in what values are considered good. I need to run an anova analysis for each test to know how the appliances compare for that test. Using the values associated only with that test. The end goal is to be able to see the f and P-values for each test in a table. 

How can I do this? 

Link to comment
Share on other sites

  • 2 weeks later...

Hello Tanner, The ANOVA tool in Spotfire is optimized for "wide" data, where it can do ANOVA test for multiple categorical variables versus multiple numerical variables and the result will be a table of F and p values for each pair of categorical and numerical variables. Once the calculation is done, you can navigate box plots of every pair by stepping through the table. You might be having difficulty because your data is "tall and skinny" I mean that all of your tests are stacked on top of each other, which sounds like what you have with Appliance column, Test column (multiple test names in this column) and Value (all of the values for all of the tests in this column). If that is the case, you could try to "Pivot" your data - this is a built in data transformation in Spotfire - the result will be a "wide" dataset, where you have Appliance column, Value for Test1 column, Value for Test2 column etc. Then you pick Appliance for your Categorical (or you could choose multiple categorical columns if you have them) and choose all of the Test columns as your Numericals. You'll get the table I think you're looking for as a result, it's not a Trellis, but a simple table listing the statistics for each combination, sorted by the most significant.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...