Where is the Scatterplot Story Analysis Useful?

• Regression Analysis - call out relationships between two measures to identify impact
• Identify Outliers - more easily understand outliers above and below thresholds

With a scatterplot story type, a user is trying to better understand the relationship between two measures. The narrative analysis adds the relationship (regression) between two measures, and if there are any groups (clusters) within the data.

Example of analytic types found in a Scatterplot Story:

Scatterplot Story Requirements

This guide will go into more detail about how Narratives for Qlik analyzes scatterplots and performs regression analysis.

Scatterplots with Two Measures

In this example, a Qlik sheet has a scatterplot with a single dimension (Product Group) and contains two measures (Back Orders and Margin Amount).

Order of the measures matters: The first measure in the sidebar is what will appear on the x-axis. The second measure will be on the y-axis. Therefore, the narrative will analyze the impact of the independent variable (Back Orders) has on the dependent variable (Margin Amount).

Scatterplots with Three Measures

In this example, a Qlik sheet has a scatterplot with a single dimension (Product Group) and contains three measures (Back Orders, Horsepower and Margin Amount).

While this scatterplot has three measures, the user will still see a two dimensional plot (i.e., just a visible x-axis and y-axis). However, the size of the third measure values is represented by the size of the bubbles in the plot.

Similar to the two dimensional scatterplot, order of the measures on the right does matter. When scatterplots are created with x,y and z-axis, analysis is typically focused on the impact that x and y have on the z-axis.

In our example, the relevant analysis is answering the question, "What was the impact each of Back Orders and Horse Power had on Margin Amount (if any)?"

If a user wants to change those independent and dependent variables, the measures on the right sidebar can be dragged and dropped into a different order to change the regression analysis performed.