# Power Bi key influencers – Complete tutorial

In this power bi tutorial, we will discuss What is Power Bi Key influencers. And also we will discuss the below points:

• When to use Power bi key influencer?
• Features of the power bi key influencer
• How to create Power bi Key influencer visual
• Interpret categorical Power bi key influencer
• Power bi key influencer interact with other visual
• Top segment in power bi key influencer
• Enable adding count feature in power bi key influencer
• Limitation of power bi key influencer visual
• Power bi key influencers no influencers found
• Power bi key influencer metric has more than 10 unique values
• What are the data point limit for key influencer?
• Power bi key influencers algorithm

## What is Power bi Key influencer?

• The power bi key influencer helps to understand the factor that drive a metric you are intrested in.
• Power bi key influncer analyzes the data, ranks the factor that matter and display them as key influencers.
• Power bi key influencer visualization as a machine learning solution to allow bussiness to leverage AI (Artificial intelligence).
• So that they can analyze their data in less time and make bussiness decision faster.

## When to use Power bi key influencer?

Let’s see when to use the Power bi key influencer in the power bi desktop.

The Power bi key influencer is a perfect choice when you want:

1. To see which factors affect the metric and being analyzed.
2. To contrast the relative importance of these factor.

## Features of Power Bi key influencer

Here we will see features of power bi key influencer.

1. Tabs: Power bi key influencer visual contain two tabs i.e. Key influencer and top segments. The Power bi key influencer tab shows you the top contribuors to the selected metric values. Where as the Top segment shows you the top segment that contribute to the selected metric value. A segment is made up of a combination of values.
2. Drop Down box: In the power bi key influencer tab, where the value of metric under investigation. You can select the metric from the drop-down.
3. Restatement: This feature helps you to interpret the visual in left pane.
4. Left Pane:In the power bi key influencer visual, the left pane contains one visual. And the left pane shows a list of the top key influencer.
5. Restatement: It helps you interpret the visual in the right pane.
6. Right Pane: In the key influencer visual, the right pane contain one visual. Whereas in the right pane the specific value of usuability shown in green and all other values are shown in blue.
7. Average line: In the right pane, the average is calculated for all possible for the selected influencer. It also applies to the values in blue.
8. Check box: This feature helps to filter out the visual in the right pane to only show values, that are influencers for that field.

## How to create Power bi Key influencer visual

Here we will see how to create a power bi key influencer visual with categorical metrics in power bi desktop using the sample data.

• Open your Power bi desktop
• Load the data using Get data from the ribbon.
• Click on the Key influencer visual from the visualization pane.
• Then move the metrics you want to investigate to Analyze field, here i am moving Rating from the Customer Table -> Rating
• Then move the field you might think that influence Analyze field(Rating) to the Explain by field, here I am moving field from the field pane these are
• Role in org
• Company size
• Subscription Type
• Theme
• Country-region
• In the below screen shot, you can see the Power bi key influence visual.
• To focus on the negetinve Rating, select the rating as Low in the What influence Rating to be drop-down box.
• This analysis runs on the table level of the field that is being analyzed and in this case it is Rating Metric.
• This metric is also defined at a customer level, each customer has given either high score or low score
• All the metrics in Explained by must be defined at the customer level for the visual to make use of them.
• All of the explanatory factor have either many to one or one to one relationship with the metric.
• Each customer assigned a single theme to their rating.
• Similarly customer from one country, holding one role in their organization and have one membership type.

## Interpret categorical key influencer in power bi

Here we will see interpret categorical key influencer in power bi by using the above Power bi key influencer visualization as an example.

Let’s discuss interpret categorical influencers for the low rating (above example), that is

• Top single factor that influences the likelihood of a low rating.
• Second single factor that influences the low rating.

### Top single factor that influences the low rating

From this example, the customer has three roles these are administrator, consumer, and publisher.

In this example, the consumer is the top factor that contributes to a low rating.

• In the left, by selecting Role in Org is consumer, Power bi shows more details in the Right pane.
• From this right pane chart, we can see 14.93% consumer given low score.
• All other roles given 5.78% low score of the time, on average.
• By deviding the blue bar by the red dotted line, you can determine consumer are 2.57 times more likely to give low score, as compared to all other roles.

### The second single factor influences the low rating

In this example, we will see the second single factor influences the low rating.

The power bi key influencer compares and ranks the factor from many different variables. So the second influencer on the list is Theme usability.

• The theme of customer review is second most important factor influences the low rating.
• The customer who commented about the theme usabilty of the product were 2.55 times more likely to give a low score to compared to consumer who commented on other theme such as relaibility speed or design.
• Between the above visual and this visual, the average, shown in the dotted line, changed from 5.78% to 11.34%.
• For the first influencer in visualization, the average excluded the customer role, and for the second influencer, the average excluded the usability of theme.
• The average is dynamic in power bi key influencer visualization, because it’s based on the average of all other visuals.
• By clicking on the Only shows values that are influencers checkbox, it will filter the visual by only using the influential visual.
• Now you can see 12 themes reduced to 4 that power bi, identified as the theme that get low rating.

## Power bi key influencer interact with other visual

Here we will see power bi key influencer interact with other visuals in power bi desktop.

I will use the above power bi key influencer visualization and I will take a slicer that will interact with the key influencer visualization.

• Lets take a slicer from the visualization pane in power bi desktop
• Add the company size to the Fields from the field pane.
• In the Power bi slicer visualization select > 50000, which will rerun the analysis and then you can see the influencer changed
• And for the large company customers, the top influencer for low rating has a theme related security.

Every time when you select a slicer, filter, or other visuals in canvas, the power bi key influencer visualization reruns its analysis in the new portion of data.

## Top segment in power bi key influencer

Here we will see the top segment in power bi key influencer. We will use the above power bi key influencer visualization as an example.

• We use Key influencer tab to asses each factor individually. Where as in top segment tab to see how a combination of factor affect the metric that you are analyzing.
• Initially top segment shows an overview of all segment that Power bi discovered.
• In the power bi visualization select the top segment tab, and you can see the six segment
• In the power bi key influencer visualization, these segements are ranked by the percentage of low rating within the segments.
• From this example, the segment 1 has 30.8% customer rating that are low.
• In the top segment, the higher the bubble the higher the proportion of low rating.
• And the size of the bubble in top segment, represent how many customer are within the segment.
• By selecting a bubble you can see the details of that segment.
• In this example, click on the top of the segment 1 bubble, you find that it is made up of relatively established customer.
• Role of org they are not publisher, so they are either administrator or consumer.
• Subscription type is primer who have low rating. And Theme is security who get low rating.
• In this group 30.8% customer gave low rating. And the averrage customer gave low rating of 11.7% of the time.
• So this segement has a larger proportion of low rating. It is 19 percentage higher than average(11.7%). And segment 1 contains 3.5% of data.

## Enable adding count feature in Power bi key influencer

Here we will see how to add count feature in power bi key influencer. We will take the above key influencer visualization as an example.

• For example, theme usability is the second largest influencer for low rating.
• There might have only been a handful of customer who complained about usability.
• Here Count feature will you to prioritize which influencer to focus, for that turn on the counts.
• To turn on the count, go to Format section in visualization pane. Expand the Analysis section And turn on the Count.
• And also you can change the type of count. If you select the relative then it will show the most amount of data, and all other count will relative to it.
• Once you turned on the count, you will see ring around each influencer’s bubble, which represents the approximate percentage of data that influencer contain.
• The more the ring circles of the bubble, the more data it contains.
• And you can see that Theme’s usability contain a small proportion of data.
• Now you can use the Sort by toggle in the left side and bottom of power bi key influencer visualization. Select the count, to sort the bubble based on the count (type is absolute).
• Based on the count the top influencer is Subscription type is primer.

## Limitation of power bi key influencer visual

Here we will see the Limitation of power bi key influencer visual.

Limitation of power bi key influencer visual are:

• The Direct query is not supported for this visualization.
• In this visualization Publish to web is not supported.
• In SharePoint Online the visualization embeding is not supported
• This visualization required .NET framework 4.6 or higher is required.
• In this visualization, live connection to Azure Analysis service and SQL server analysis service is not supported.

## Power bi key influencers no influencers found

Here we will discuss Power bi key influencers no influencers found an error in power bi desktop.

This error ‘no influencer found’ occur, after adding fields to the Explain by:

• You were included the metric, you were analyzing in both the Analyze and Explain by field in field format. Then remove the metric from the Explain by.
• If your explonatory fields have too many categories with few observation. Then in this situation makes it hard for the visualization to determine which factors are influencers.
• And also it’s hard to generalize based on the few observation.
• If you are analyzing the numeric field in key influencer visualization, then you may want to switch from categorical analysis to continuous analysis in the Fromatting pane-> Analysis section.
• Your explonatory factors have enough observation to generalize, but the power bi key influencer visualization didn’t find any meaningful correlation to report.

## Power bi key influencer metric has more than 10 unique values

Here we will discuss power bi key influencer ‘metric has more than 10 unique values’ error occur in power bi desktop.

The reason behind ‘Metric has more than 10 unique values. This may impact the Quality of the analysis‘ error happen:

• In the key influencer visualization, the AI can analyze the categorical field and numerical field. The categorical fields, may be Yes or No and customer satisfication is may be high, low or medium.
• So the increase the number of categories to analyze, means per category there are fewer observation.
• So in this situation it is harder for the Key influencer visualization to find the patterns in the data.
• When we analyzing the numeric field, we have a choice between treating the numeric fields like text, in this case you can run the same analysis as you do for categorical Analysis/for categorical data.
• If you have lots of distinct value, then it is recommended to switch to the Continuous Analysis.
• By switching to continuous Analysis, you can infer patterns when number increase or decrease rather than treating them as distinct value.
• To get stronger influencer, it is recomended you to group similar values into single unit.

## The data point limit for Power bi key influencer

Here we will see what are the data point limit for power bi key influencer.

• In power bi, when we run the analysis on a sample of 10,000 data points.
• In the power bi key influencer visualization, in one side it shows bubbles of all the influencer that were found.
• Whereas the other side, column chart and scatter plots, it abide by the sampling startegies.

## Power bi key influencers algorithm

Here we will see the Power bi key influencer algorithm by ML.NET. And also we will see

• How power bi key influencer algorithm works ?
• How power bi key influencer algorithm works for categorical key influencer?
• How power bi key influencer algorithm works for Numeric Key influencer?
• How power bi key influencer algorithm works for calculating top segments

Introduction to Power bi Key influences algorithm

• When a user picks KPI(key performance indicator) to analyze, the power bi key influencer visualization uses Machine learning algorithm by ML.NET.
• It helps to figure out what matters the driving metrics, as well as to find intresting segment for further investigation in power bi.
• Power bi key influencer analyzes the users data, contrast the relative importance of these factor, and ranks the factor that matter.
• And then display them as power bi key influencer and top segment for both the categorical and numerical values.

How does the power bi key influencer algorithm work?

Let’s see how the power bi key influencer algorithm works.

• When a user add a columns to the power bi key influencer visualization, a flow is triggered in which training data sent to the Analysis service (the database engine behind Power bi).
• Where as the Analysis service runs ML.NET to train the Machine learning(ML) models, and results are returned.
• And the Machine learning model is trained whenever a user updated the selected features.
• The goal is to perform the analysis in a few seconds, by enabling an interactive experience.

How does the power bi key influencer algorithm work for categorical key influencers?

Now we will see power bi key influencer algorithm works for a categorical key influencer.

• In power bi key influencer, the categorical metrics can includes things like rating and ranking
• The power bi key influencers, uses the ML.NET to run a logistic regression to calculate the key influencer.
• Whereas the logistic regression is a stastical model that compares different groups to each other
• In this example the metric is rating. Here if user want to see, what drives the low rating, the logistic regression looks at how customers who gave low score differ from the customer who gave high score.
• If there is multiple categories, like high, neutral, and low scores, you look at how the customer who gave low rating differ from the customer who didn’t give low rating.
• In the data the logistic regression searches for patterns and looks how customer who gave low rating may differ from the customer who gave high rating.
• Also the logistc regression consider how many data points are present.
• When there are not enough data points to inffer a pattern.
• A stastical test is done, which is used to determine wheather a factor is considered an influencer, it is also known as Wald test.
• The power bi visuals uses a p-value of 0.05.

How does the power bi key influencer algorithm work for Numeric Key influencer?

Let’s see power bi key influencer algorithms work for a numeric key influencer.

• In power bi key influencer the numeric metric can include things like price or sales number.
• The Power bi key influencer visualization uses ML.NET to run a linear regression to calculate the Key influencer.
• Whereas a linear regression is a stastical model, that looks at how the outcome of the field you are analyzing changes based on the explainotary value of the visualization.
• The linear regression is also consider the number of data points in visualization.
• When there are not enough data points available infer to pattern, this factor is not considered influential.
• The stastical test is done, to determine wheather a factor is considered as influential, which is also known as Wald test.
• The power bi key influential visualization uses p-value of 0.05 to determine the threshold.

How power bi key influencers algorithm works for calculating top segments

Let’s see power bi key influencer algorithms work for calculating top segments.

• In power bi key influencer, the top segment tab shows the top groups that contribute to select the metric value.
• In top segments, a segment is made up of a combination of values.
• In the power bi key influencer visualization uses ML.NET to run a decision tree to find intresting subgroups.
• Whereas the decision tree objective is to end up with subgroup of data points, that is relatively high in the metric you are intrested in.
• The decision tree takes each explonatary factor of the visualization and tries to reason which factor gives it the best split.
• Once the decision tree does a split, it takes the subgroup of data and determines the next best split for the data .
• And after each split, it considered weathers it has enough data points for the group to be representative enough to infer a pattern from wheather or it is an anomaly in the data and not a real segment.
• And another stastical test is applied to check for the stastical significance of the split condition with p-value is 0.05.
• When the decision tree finishes running, it takes all the splits and then creates a filter in power bi.
• In the visual the combination of filter is packaged up as a segment.

You may like the following Power Bi tutorials:

In this power bi tutorial, we discussed power bi Key influencer. And also discussed the below points:

• When to use Power bi key influencer?
• Features of the power bi key influencer.
• How to create Power bi Key influencer visual.
• Interpret categorical Power bi key influencer
• Power bi key influencer interact with other visual
• Top segment in power bi key influencer
• Enable adding count feature in power bi key influencer
• Limitation of power bi key influencer visual
• Power bi key influencers no influencers found
• Power bi key influencer metric has more than 10 unique values
• What are the data point limit for key influencer?
• Power bi key influencers algorithim