Correlation Matrix Not Available

Options
sdfungayi
sdfungayi Dataiku DSS Core Designer, Dataiku DSS & SQL, Dataiku DSS ML Practitioner, Registered Posts: 19 ✭✭✭✭✭

I want to perform correlation matrix analysis on a dataset but all the fields are greyed out and not available, making it impossible to do the analysis.

I was expecting to be able to select fields for use in the analysis, but DSS greys out all the fields immediately I enter the "Correlation matrix" window.

What could I be missing?

See attached image

Fields not available.JPG

Best Answers

  • Clément_Stenac
    Clément_Stenac Dataiker, Dataiku DSS Core Designer, Registered Posts: 753 Dataiker
    Answer ✓
    Options

    Hi,

    The Statistics capability in DSS relies on the storage type of the columns to know if they are numerical or categorical. Only numerical columns are eligible for correlation matrix. At the moment, all your columns are categorical (i.e. "string")

    You have two options:

    * Use a prepare recipe to create a new output dataset with types infered from the input
    * Or go to the Settings of the dataset, then go to Schema, click on "Check now" then on "Infer types from data" in order to compute storage types that match your data

  • sdfungayi
    sdfungayi Dataiku DSS Core Designer, Dataiku DSS & SQL, Dataiku DSS ML Practitioner, Registered Posts: 19 ✭✭✭✭✭
    Answer ✓
    Options

    I tried both options. They work perfectly.

    Thanks

Setup Info
    Tags
      Help me…