Power BI – Key Influencers chart

In this article, we are going to see Power BI key influencers charts.

We will be discussing the following topics and their basic implementation in the Power BI desktop.

Key influencers Charts

This is a Power BI Artificial intelligence visualization that helps us to understand the factors that drive a metric of your interest. It thoroughly analyzes data and ranks the factors that matter which are commonly known as key influencers. It helps in showing the top contributors and also shows the top segments (combination of values) for some selected metrics.

When to use key influencer’s visualizations

• These are used when the user wants to know the factors that matter the most for the selected metric.
• Analyze a field explained by other data fields.

You can refer to Power BI interactive dashboards for easy implementation of the following charts.

Features of key influencer’s visualizations

Let us understand the features of key influencers’ visualizations by using the following example image with colored boxes. Each color represents a different feature.

1. Tabs: The red colored box represents tabs as “Key influencers” and “Top Segments” which can be selected for switching between views. The “Top Segments” tab shows the top segments (combination of values) contributing to the selected metric and “Key influencers” are top contributers for the selected metric.
2. Dropdown box: The green colored box represents the dropdown box having all the values of the selected metric. The key contributers are analyzed by selecting any value from the dropbox.
3. Restatement: The yellow colored box helps in intepreting the visual in the left pane.
4. Left pane: The blue colored box shows the left pane showing the top key influencers.
5. Restatement: The yellow colored box helps in intepreting the visual in the right pane.
6. Right pane: The blue colored box shows the right pane with the values of the top key influencers for detailed analysis.
7. Average line: The average is calculated for all possible values for except usability (which is in the light blue column selected influencer). The average calucluation is applied for all dark blue-colored columns.
8. Check box: The orange yellow colored box in the bottom right filters out the visual in the right pane to show values that are the only influencers for that field.

Key influencers are the factors that affect or drive any outcome. There are two types of outcomes.

1. Categorical key influencers
2. Continuous (numerical) key influencers

Interpret categorical key influencers

Categorical key inflencers are the key factors or indepedent variables which makes the outcome to change depending on it. Let us understand with the following example.

Consider analyzing “Loan” explained by other fields dragged to “Explain by” field. The setting of the visualization pane is given further while scrolling down in the article. We are just showing images to understand the 2 types of key influencers initially.

Factors that influences the likelihood of increase in loan intake

We have kept the dropdown to “Increase” to see at what influences loan to increase.

Top single factor is “Interest_rate“. The “Interest_rate” of 13.12 or less is the top factor that contributes to a increase of “Loan“. More precisely, if “interest_rate” is 13.12 or less then your “loan” increases by 121.3 times.

When “Salary” increases, the “Loan” also increases. When the “Salary” increases by 48663, the “loan” also increases by 61.6 times.

People with age 46-70 are 26.61 times more likely to have a increase in “loan” intake.

Note: Select the “Only show values that are influencers” check box to filter by using only the influential values. Refer to the above image in the bottom right pane.

Interact with other visuals

Every time the user select a filter or slicer or any other visual on the canvas, the key influencers runs again as a new analysis on the new set of data.

For example, you can move “Quarter” into the report and use it as a filter. Use it to see the analysis for the third quarter of any year.

Key influencer chart after using the filter field

Consider the chart is filtered for the third quarter of any year. For the third quarter, the loan is more to likely to increase when the interest_rate is 13.56 to 13.8.

Similarly the second and third top influencer also gets changed a little after the filter. The user can investigate further for different quarters and see that the interest_rate and other factors gets changed for increase or decrease in loan intake.

Interpret continuos key influencer

We have seen how categorical fields influence the increase in loan intake. There are also continuous or numerical factors like age, price, size,population count in the “Explain by” field.

When total “units” increases, the “Sale_amt” also increases.

Let’s look at what happens when “SaleTeamSize” is moved from the table into “Explain by” field. We can notice a change in the above image as follows. When the “SaleTeamSize” is added, the “Sale_amt” gets increased from 9.32 times to 15.8 times on average.

Another example is as follows

As “Total_Published” articles increases, the likelihood of “Expected” number of articles also increases.The scatter plot in the right pane shows the average number of expected articles for each value of “Total_Published“. It highlights the slope with a trend line.

Now let us understand the Key influencers chart in detail with the help of excel datasets.

DataSet used:

The dataset used is “SaleData.xlsx “. Upload the dataset in Power BI and refer to the dataset to follow along with the below-given sections of the article.

SaleData.xlsx: A sample screenshot of the dataset is taken to get an idea of the data fields for analysis of various metrics.

Data: We will be working with “SaleData.xlsx” data with data fields as shown in the above image. The major variables used to show the charts are as follows.

• Sale_amt: The total sale amount for various items of different brands.
• Units: The total number of units sold for any given item of any brand.
• Unit_price: The price of any item.
• Region: The different regions are “east”,”west” and “central” where the sale is handled.

There are other data fields also which can affect any given metric. The user can make use of other fields as per the need or requirement for the analysis or prediction.

Load Data in Desktop:

Open Power BI Desktop

Click the “Get Data” option and select “Excel” for data source selection and extraction.

Select the relevant desired file from the folder for data load. In this case, the file is “SaleData.xlsx”. click the “Load” button once the preview is shown for the Excel data file.

The dashboard is seen once the file is loaded with the “Visualizations” Pane and “Data fields” Pane which stores the different variables of the data file.

In the “Visualizations” pane, we can select the chart type needed for our visual representation of business data. In the following image, the red colored square represents the “key influencers” chart that can be dragged for the “Report” view for further process.

Initial Key Influencer Visual: When the red-colored icon is dragged to the main visualization pane, we get the following view initially.

Key Influencers Chart in Power BI

Steps for any metric analysis:

• Move the metric you want to investigate into the “Analyze” field of the “Visualizations” pane from the “Data” pane.
• Move fields that you think might influence the field in the “Analyze” field into the “Explain by” field. You can move as many fields as you want. Refer to the following image for initial setting.
• Leave the Expand by field empty. This field is only used when analyzing a measure or summarized field.

Example 1

Analyze Sale amount by Brand and Unit Price.

We will analyze the important metric “Sale_amt” explained by brand and unit price for any item sold.

Note: The analysis runs on the table level of the field that’s being analyzed.

Output: This is the output seen in the main report view pane.

It shows the key influencers for the total sale amount “Sale_amt”. It predicts the following.

• When the unit_price of any item increases, the sale amount also increases. On click of the blue circle or hovering over it will give full information regarding the key influencer or metric.

Output:

Key influencers: When the total number of “Units” and “Unit_price” of any item increases, the average “Sale_amt” also increases by a certain value as shown in the given output image. Keep hovering on the right hand side blue and grey colored circles to get a detailed view or information for any prediction.

This video shows the key influencers and analysis for better understanding of the Power BI KI charts. The user can change the dropdown to see the changing effect of data and influencers. The “Count of SaleData” shows the total number of rows in that particular setting.

The following shows another explanation when the total “Units” of any item increases in the sale process.

Example 2

We can work out the similar steps with different datasets or excel sheets. The following screen shows various columns of “DataSet.xlsx” which stores data for loan, interest rate and salary information of people with different ages and gender. The loan is taken for various purposes like “Home”, “Education”, “Travel”, “Personal” and “Other” reasons in different time line like “Year”, “Quarter” and “Month”.

We can make analysis using the Power BI key influencer chart according to the needed metric.

DataSet.xlsx:

Analyze Loan By Interest Rate: The visualization pane setting can be done by following the above steps as explained earlier in example 1 of this article. Let us analyze various “Loan” taken with different “Interest_rate”.

Output: From the data table, we can see that there are various interest rates on the loan like 19.8, 19.2,18.8 % and so on. When the Interest_rate is 13.12 or less, the loan increases by 121.3k on average. The light-blue colored column shows the values that are influencers. The x-axis shows various interest_rate ranges and y-axis shows the values for the average of loan and its increase.

Analyze Laon by Salary and Date: Let us analyze various loans taken based on salary of people and various timeline. Set the “Visualizations” pane as shown below by moving required fields from the “Data” pane to the relevant “Analyze” and “Explain by” fields. The “Date” can be in the form of “Year”, “Quarter”,”Month” or “Date” (“dd/mm/yyyy”) format. Select date format as per the requirement.

Output: The output shows that when the salary increases, the loan also increases.

Analyze loan by age: From the above predictions of key influencers, we can analyze different fields explained by different fields.

The loan taken by people seems to increase in the age group of 46 to 70 on average.

Analyze gender by average salary and average remaining debt:

Visual pane: Set the relevant fields as given below.

Output: The gender can be changed among “F” (female) and “M” (male) in the left side dropdown box. The following key influencers are given based on the gender. The gender is more likely to be “F” when the average of salary is 20510 and so on. For detailed information, change or click the tab to “Top segments” and click and hover on various circles.

Analyze Loan by average interest rate and age: We analyze loan based on average of interest rate and age.

Visual pane:

Output: Loan is more likely to increase when the average interest rate is 13.12 or less within the age group of 46 to 70.

Top segments: Click the other tab of “Top segments” for detailed information of loan to be “high” or “low”.

Analyze Loan by average salary and age: We analyze loan by average salary and age.

Output:

Top Segments: See the below output to predict “When the loan is more likely to be low?”

Analyze Purpose of loan by year, average salary, remaining debt and interest rate:

The following key influencer chart shows the various purposes for which the loan has been taken based on salary, remaining debt to be fulfilled and interest rate along with the date and year.

Output: The various purpose of loan like “Home”, “Education” and so on are in the dropdown box which are selected by the user as per the requirement analysis.

Example 3

We are showing another example with the following data set.

techPublishedDataSet.xlsx:

Follow the steps as explained in the initial sections of the article to navigate and load data of excel sheet file in the Power BI desktop.

Analyze Total articles published by Technology and sum of expected articles to be published:

Output: The following shows the total number of articles which are finally published based on some particular technologies like “PHP”, “AI” and so on explained by total number of expected articles for publish.

Note: While creating the key influencers chart, if the user encounters a error showing “No influencers found.Try adding some more fields into “Explain by”, it occurs due to the following reasons.

1. When the fields are added in the “Explain by” field, but no influencers are found.
2. The metric to be analyzed is dragged both to “Analyze” and “Explain by” fields.
3. The fields in the “Explain by” section have too many categories with very few observations.
4. Sometimes many observations are present but there are no correlation in the visualization to report.