Scatter Plot Matrix Examples

This workflow explains how to use the Scatter Plot Matrix node and provides an interactive view with the node inside.

You can easily download and run the workflow directly in your KNIME installation. We recommend that you use the latest version of the KNIME Analytics Platform for optimal performance. It can also be deployed as a Data App in KNIME Business Hub.

We begin by using the CSV Reader node to load "The Boston Housing Dataset," which has already been saved in the workflow data area for easy execution on your local installation.

This dataset is ideal for demonstrating the Scatter Plot Matrix node. It has two independent variables, "median home value" and "nitric oxide concentration", which are affected differently by other variables. All dataset features are numerical.

To demonstrate how to customize chart colours, we created a categorical column with the Auto-Binner node and Rule Engine based on the "median home value" column. This resulted in a "clustered" column with each row assigned to a bin labelled Low, Medium-Low, Medium-High, or High value.

Using the Color Manager node, we can customize the colour of the new categorical feature "median home value [Binned]" to better visualize the potential relationship between variables in the matrix view.

In the data application, we follow the steps in the "Enriched Configuration" branch to create a categorized binned column. We have the option to choose between two features to apply colour: "median home value" or "nitric oxides concentration".

Click on the "nitric oxide concentration" option in the view. The colour scheme shows the negative correlation between distance from Boston and nitric oxide concentration in the High-value bucket, meaning properties near the city have worse air quality.

Switching to the "median home value" colour shows quickly the direct correlation between it and the "% lower status of the population."

External resources

• The Boston Housing Dataset - Kaggle