Using key influencers in Power BI
1. Using key influencers in Power BI
In this final video, we’ll dive into the basics of the Key influencer visualization native to Power BI. As mentioned in the previous video, this allows us to understand the influence a chosen explanatory variable has on a target outcome. Said in more statistical modeling terms, how much the explanatory variable explains the variance in the outcome. Sticking with our deep dive into monthly income, on a new page, I’ll create a new Key influencers visualizations for MonthlyIncome. For our “Explain by” variables, let’s start simple with TotalWorkingYears. The result shows TotalWorkingYears as a key influencer on average monthly income. Specifically, as total working years increases by 7.66 year, average monthly income will increase by $3,420. There is also a scatter plot accompanying this information, which looks similar to the one created in a previous exercise. Interesting. But the power of this visualization starts to reveal itself with multi-variable analysis. So, I’ll add Gender, PerformanceRating, RelationshipSatisfaction, YearsAtCompany, YearsInCurrentRole, YearsSinceLastPromotion, YearsWithCurrManager, JobRole, Department, and Education. There are multiple key influencers – as YearsAtCompany goes up 6.51 years, average monthly income increases only $148; the “Manager” job role leads to a $11,760 increase in average monthly income. Clicking on the circle associated with the Manager key influencers will show a chart with further breakdown of average monthly income by role. For simplicity, I’ll remove the variables that aren’t key influencers at the moment. Right now, the visualization is automatically looking for key influencers in the increase of average monthly income. We can change this to look at what influences the decrease by clicking on this dropdown and selecting decrease. We see being a Sales Representative leads to a $3,840 decrease in average monthly income. I’ll switch back to “increase”. The Key influencers visualization provides further information known as Top Segments. Here, Power BI automatically clusters together similar observations when the target variable is most likely to be high or low since it is continuous. For a binary categorical variable, it will cluster by true or false. Here, we see two segments. The first is of 25 observations and has an average monthly income of $18,170. The second has 37 observations and an average of $13,140. Clicking on each circle will reveal characteristics of each cluster. The first cluster is characterized by the Manager job role, more than 20 years of working experience. The average income is $12,150 higher than the overall average. The second cluster are non-manager individuals but still with more than 20 years of work experience. Their average is $7,120 higher than the overall average. Again, we can switch to segments where the average monthly income is more likely to be low. Here we see there are more segments to explore. But for now, it’s your turn to use the Key influencers visualization to analyze the influential variables on the price of AirBnb listings.2. Let's practice!
Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.