March 28, 2019

How to Use Power BI’s AI Visual: Key Influencers

Post by: Marcus Radue

Microsoft has recently released a new Key Influencers visual in their February 2019 release of Power BI. This visual is part of Microsoft’s roadmap to continue to advance the Artificial Intelligence (AI) integration and features within Power BI. Microsoft has already introduced other AI features within Power BI such as Quick Insights and Natural Language Querying for live Q & A.

The Key Influencers visual is another way for your report developers to leverage Microsoft’s machine learning technology and gain great insights into your data.

Preview State

Note: The Key Influencers visual is currently in a preview state and with that comes some limitations. Those current limitations include:

  • Analyzing metrics that are aggregates/measures
  • Consuming the visual in Power BI Embedded
  • Consuming the visual on Power BI mobile apps
  • RLS support
  • Direct Query support
  • Live Connection support

Use Case

As we observe the NCAA college basketball tournaments during this exciting month of March Madness, I deemed it appropriate to use a college basketball dataset for my sample Power BI report. That dataset includes all the men’s division 1 college basketball teams that are eligible to receive an NCAA tournament bid from the years 2003 – 2018, along with their regular season and NCAA tournament records and statistics for those years. With the Key Influencers visual, I was able to gain insights to the criteria that may have factored into the NCAA tournament seeding over those years. Let’s look at the data trends and factors I was able to gain that may have influenced the tournament seeding.

Download the sample Power BI report and dataset.

Key Influencers Visual Example

Before demonstrating how the Key Influencers visual works, you will need to make sure you have the latest release of Power BI desktop installed on your device. If you have Power BI desktop installed via the Microsoft app store on a Windows 10 machine, then Power BI will update automatically. If you have an earlier Windows version, then you can get the latest version here.

The second thing you will need to check is that your preview features are checked on in your Power BI desktop settings. You can do that by opening a new Power BI desktop file, going to File > Options and Settings > Options. I have all my preview features checked to make sure I always have the full functionality of Power BI, but you only need to have the Key Influencers visual checked.

Once you have done that, you can load the sample dataset I have provided and select a new Key Influencers visual onto your report page.

Here is my finished Key Influencers visual used to describe the factors that may have went into deciding the NCAA Tournament seeding spanning the years between 2003 and 2018. Within the visual you will have two tabs: Key Influencers and Top Segments. Notice you can change the value the visual is analyzing within the drop-down menu at the top of the chart. You can also use other slicers on your report page to interact with your Key Influencers visual.

Key Influencers Tab

Note the field I chose to analyze and then the corresponding fields I chose to explain. Because of a current preview limitation of not being able to analyze by aggregations, I chose to analyze the NCAA Tournament seeds. Some of the trends you can see from all the number 1 seeds include:

  • The team with the most number 1 seeds from 2003 – 2018 is Kansas. The closest 3 teams behind Kansas are North Carolina, Duke, and Kentucky.
  • To gain a number 1 seed you must have a high number of regular season wins. This insight is stated from the “RegWins goes up 8.21”.

If I scroll down in the visual, I can see that the top 3 conferences that have received number 1 seeds from 2003 – 2018 are the Atlantic Coast, Big East, and Big 12 conferences. Notice that clicking a different key influencer in my visual changes my graphical display.

Top Segments Tab

By clicking the Top Segments tab, you get taken to a new view that shows the different groupings of the Key Influencers and the impact they had on the selected value – in this case, tournament seed.

Notice the different groupings created from the Key Influencers. Clicking a segment drills deeper, showing the statistics derived from that segment grouping.

You can see that segment 1 shows the number 1 seeds that had 5 or fewer regular season losses and 29 or greater regular season wins. This was the largest segment of number 1 seeds at 67.5%. You can infer that a major factor to gaining a number 1 seed was having a high win and low loss total on a team’s regular season record. This seems like a logical explanation – that the best team records receive the best tournament seeds. While it may be a simple example, it proves that the Key Visualizers visual is doing the correct analysis on the dataset.

The “Learn more about this segment” goes even further down, giving deeper insights on the segment grouping.

Conclusion

While the new Key Influencers visual may not be appropriate for every business use case, its value is undeniable for Power BI report creation. As a developer, it’s exciting to be able to leverage machine learning within an interactive report environment to gain deeper insights about your data! This visual also demonstrates Microsoft’s continuing investment into Power BI and its machine learning features. I encourage you to STAY TUNED to Microsoft’s monthly Power BI updates and development roadmap.

If you found this blog interesting or want to learn more on how Power BI can help you or your business, let’s talk about your goals.

New call-to-action

Subscribe to our Newsletter

Stay informed on the latest technology news and trends

Relevant Insights

Should You Disrupt Yourself to Accelerate Digital Transformation?

It has been interesting to watch Microsoft transition from a company that makes its money via licensing to one that...

Cybersecurity Myth Busted: Tools Are the Solution

When thinking about security, people often gravitate towards implementing various security tools, solutions, or products. If you bring up a...

Time to Reconsider MP-BGP EVPN for Your Datacenter Network?

VxLAN was defined in 2014 by RFC 7348 and has been used as a component in several SDN (software defined...
X