Understanding the factors that drive a metric with Power BI’s Key influencers visual

The purpose of the key influencers visual is to assist in finding individual factors that drive an outcome.  An outcome can be Categorical or Numerical (Continuous). An example of a Categorical outcome is something like customer reviews that are positive or negative. And a Numerical outcome is something like an Income variable before summarising it into Income levels.

To calculate and model the Key influencers, Power BI makes use of Regression analysis. Regression analysis is a statistical method. This method calculates how the dependent variable (outcome of the field you are analyzing) changes based on the independent variables (explanatory variables).  So, the regression is trying to find the correlation between variables.

For this blog, we will look at the factors affecting the level of income. We’ll be using the AdventureWorks dataset. Depending on the dataset and sample size you use, you might get different results.

Let me take you through the steps of adding the visual. Then we’ll dive into a bit of the interpretation as well:

(Step 1) Select the Key influencers icon

Key influencer visualization


(Step 2) Select the field for analysis and fields for explain by

Factors that might be influential to the level of income are; age, level of education, marital status, and even gender.

Key influencers analyse incomelevel by education level and marital status

(Step 3) Features of the Key influencers visual

From the visual above you should note the following:

  • There are two tabs to switch between views. Key influencers show the top contributors. Top segments show the top segments that contribute to the selected metric value.
  • A dropdown box for what influences income level currently set to Average.
  • Left Visual Restatement that helps to interpret the visual in the left pane: “When the likelihood of income level being average increases”.
  • The Left pane shows a list of the top key influencers.
  • Right Visual Restatement: “Income level is more likely to be average when Occupation is…”
  • The right pane contains one visual. In this case, it is a column chart. The column chart displays values based on the selected key influencer in the left pane. Currently selected (as indicated by the aquamarine colour) is “Occupation is Professional”.
  • The average line in the bar chart is calculated for all values other than the selected influencer.
  • The Check box in the right pane is there for the option of only showing values that are influencers for that field.
(Step 4) Analysing the metric: Key influencers

In the image below, we’ve changed the dropdown from “Average” to “High”, to look at what influences Income level to be high.

Key influencers of high income level

From the column chart in the above image, you will note that the Top single factor, is Current age. Being between the age of 47 and 90 is the top factor that contributes to a high-income level. More precisely, if your current age is between 47 and 90 then your income level is 11.96 times more likely to be high.

The second single factor that influences income to be high is Total children:

Second single factor that influences income level to be high

The second influencer under selections has nothing to do with the Current age influencer. People with 2-3children are 1.68 times more likely to have a higher income level.

In this blog, we won’t go into each individual contributor. However, the same type of analysis goes for Marital status and level of education.

(Step 5) Analysing the metric: Top segments

The Key influencers tab is used to assess each factor individually. The Top segments tab is used to determine how a combination of factors affects the metric being analysed.

The Top segments tab, shows an overview of the segments as discovered by Power BI:

Top segments Power BI

There is 5 segments in this instance. The segments are ranked by the percentage of high levels of income within the segment. The higher the bubble the higher the proportion of high-income levels. The size of the bubble represents how many individuals are within that segment.

When selecting Segment 2, we see the following:

Segment selection form Key influencers Power BI

16.5% of people indicated a higher level of income. On average a higher level of income was indicated 10.2% of the time. So, this segment has a larger proportion of high levels of income. It’s 6 percentage points higher. Segment 2 also contains approximately 5.3% of the data.

Each of the other segments is analysed in a similar manner. However, for simplification, we won’t be diving into each of those.

In Summary

The key influencers’ visual is a very powerful visual. It allows the user to delve deeper by finding correlations within the dataset. For this blog, we’ve interpreted mostly categorical variables and a categorical metric. It should be noted however, that there are various other ways to interpret measures, continuous metrics, and influencers.

If you would like more detail pertaining to the types of variables that can be analysed or if you’re looking for troubleshooting tips. Then please visit the official Microsoft site:https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-influencers

For other interesting and informative blogs, you can also visit our blog site at: https://databear.com/blog/

Our training options:  https://databear.com/power-bi-training/