Analyze changes in an answer

Contextual change analysis reveals what changed and why between two data points. Examine insights by attribute to understand the contributing factors.

Before you begin

You can only run change analysis on these chart types:
  • Area
  • Line
  • Column
  • Bar
  • Line-column
  • Donut
  • KPI charts with sparklines

About this task

Change analysis compares two data points by listing their underlying attributes and identifying which attributes contribute to significant change.

Procedure

  1. Open an answer or view it in a liveboard.
  2. Optional: For KPI charts, you can quickly analyze the latest two data points by selecting the percent change label. If you do, skip step 3.
  3. Press and hold Command (Mac) or Ctrl (Windows or Linux).
    1. Select two columns or data points.
    2. Right-click on one of the selected points, then select Run change analysis.
    The Change analysis opens.
  4. By default, Explore analyzes the five most relevant attributes, based on what it learned from your past activity. To analyze different attributes, select Customize attributes, then select attributes from the list.
  5. To further analyze a specific value, right-click on an attribute value in the change analysis, then select Analyze "<value>".
  6. To save the change analysis of a single attribute, select More , then either Pin to a liveboard, Make a copy, or Download.

Example

The following change analysis is for an answer that counts the total number of open claims.

The Loss Location State attribute has statistically significant outlier values, which are shown in a chart. A summary explains that those values account for a significant percentage of the increase in number of claims.



Another attribute, Claim Tier, doesn't have statistically significant outlier values. A summary explains that the attribute doesn't significantly contribute to the change in number of claims.



Note: Sometimes attribute values contribute to more than 100% of the total change. This means that some attributes contribute to a positive trend while others contribute to a negative trend.