community
cancel
Showing results for 
Search instead for 
Did you mean: 

Alteryx designer Discussions

Find answers, ask questions, and share expertise about Alteryx Designer.

logistic regression error

Meteor

Hi-I got the most recent version of Alteryx recently. I had been able to build simple logistic models in the prior version but I'm getting this error this time around and I don't understand it (I don't know R code). My outcome variable is most certainly binary (0/1) but Alteryx is telling me the target variable must have only two unique values. It does! All of the independent variables are either scaled questions (1-5 response scale) or dummy variables I set up in advance. I used the select tool to make sure all variables were numeric/double. Any idea what this error means? Thanks!

Alteryx
Alteryx

Hi @jbuszin

 

I am sure you have checked the unique values in the promoter field, but adding a unique tool and selecting this field will give you clearest view on if this is indeed the case. 

 

I would also recommend adding a data cleansing tool before the logistic regression to make sure there are no nulls values in your fields, this can cause errors within the R code. 

 

Best,

 

Jordan Barker

Solutions Consultant

Meteor

Thanks for your reply, Jordan. Unfortunately the solution didn't fix anything. I had already used a summarize tool to make sure my dependent variable only had 2 unique values (0 and 1), which it did. I added the unique tool anyway and now my "promotor" variable, which I know has just two unique values, doesn't even appear in my "target variable" drop-down list. That's weird because I've used the select tool prior to make sure the variable is a double/numeric.

 

I also added a data cleansing tool as you suggested. Same errors as before. 

 

-Justin

Meteoroid


Justin - any luck solving this issue? I'm having the same problem.

Highlighted
Alteryx
Alteryx

HI @macd279

 

Could you check that your target variable is a string?

Meteor

considering this was over a year ago, I don't know. I assume this was figured out but unfortunately I don't remember what the solution was

Meteoroid

No worries, thanks for the response.

Meteoroid

Hi Dave, yes dependent variable is string. Also checked for nulls and special characters and all seems well. I'll keep digging. Thanks!

Labels