Overview

The Correlation node is a statistical-plot node in a Weave Workflow that measures the relationship between selected columns and renders the result as a correlation chart. Each pair of columns is scored by a correlation coefficient and displayed as a heatmap cell.

The node sits under the Statistics tab, distinct from the transformation nodes, and produces a visual output rather than a transformed table. The node applies to the columns selected in the Correlation popup.

Note: “Workflow” is the in-UI term for what some Weave documentation calls a pipeline. This reference uses “Workflow.”

When to use it

  1. Measuring how strongly two numerical columns move together.
  2. Scanning many columns at once for strong positive or negative relationships.
  3. Identifying redundant or highly related fields before modeling or further analysis.
  4. Exploring data relationships visually via a heatmap.

Coefficient interpretation

Coefficient rangeInterpretation
Near +1Strong positive relationship: the two columns increase together.
Near 0Little or no linear relationship.
Near -1Strong negative relationship: one column increases as the other decreases.

Configuration

StepDescription
SelectOpens the Correlation popup listing the columns in the upstream data.
Select AllSelects every column in the popup.
Column checkboxesChoose the columns to include in the correlation.
ApplyCommits the column selection.

Key behaviors

Statistical-plot node. Correlation is found under the Statistics tab, separate from the transformation nodes, and outputs a visualization rather than a modified data table.

Pairwise coefficients. The node computes a correlation coefficient for each pair of selected columns and presents them as a heatmap, with color and value encoding the strength and direction.

Numerical scope. Correlation is measured between numerical columns; the result reflects the numerical selections.