Column Name Missing % #Distinct Type Minimum Maximum Mean

Click a row to exclude the column from the analysis.

Select a feature of interest:
Select a data point label (for scatter plots):
More Settings

This heatmap indicates a p-value in which the null hypothesis is that the two columns are independent and not correlated.

  • A high p-value indicates the columns are independent.
  • A low p-value indicates the columns are correlated.
  • N/A indicates the test could not be performed (for Chi-square typically the expected counts are too low).

The statistical test performed is dependent on the type of data of the two features of interest:

  • Categorical vs Categorical: Chi-square test
  • Numeric vs Numeric: Pearson correlation
  • Categorical vs Numeric: ANOVA

Missing values are excluded from the p-value calculations.

Select a projection algorithm:
Label for graph legend:
Label on graph:
Missing data:
Select the feature to predict:
Select a prediction algorithm:
Missing data:
Class weights: