Microsoft Power BI Releases Game Changing Update: Easy AI-Powered Insights
The beauty of Microsoft Power BI as a platform is that it makes analytics accessible for people who aren’t data scientists. With the latest version of Power BI, the February 2019 release, Microsoft has taken that a step further and now Artificial Intelligence, Machine Learning and Advanced Analytics are accessible – even for beginners.
When I work with our customers, they tend to ask me what Artificial Intelligence is and what it can do for them. The new Key Influencers Visual from the February 2019 Power BI release is the perfect example of how they could use it – and why they should.
What is the Key Influencers Visual?
The Key Influencers Visual is a new report feature from the February update for Microsoft Power BI that shows you the key influencing factors of a certain figure. This is by far the best new update. It answers insights related questions, for example, if you wanted to know about customer churn you could find out, “What are the influencing factors as to why customers leave?”
It’s the first AI-powered visualisation to be introduced to Power BI, using machine learning in the background to deliver vital insights, so it’s ground-breaking. There’s usually a barrier of complexity or cost between simple reporting on your data and getting into the Machine Learning and Artificial Intelligence analytics. This feature is removing that barrier, it introduces people to the concept that Power BI will tell you things about your data that you may not know or even have been able to find out.
How Does It Work?
You feed Power BI the relevant attributes, variables and fields that would have any influence on a figure, let’s say your customer churn from last year. It trawls through all of the numbers, runs them through its model, and essentially outputs an indication of the factors that most impacted your customer churn. So, if your company lost customers last year – what contributed to that? Is it based on location, product, unit cost, or something else? That’s what this feature helps you to find out.
Power BI’s visuals for this feature are good at illustrating the insights you need. It uses a drag-and-drop interface to make it accessible. There’s lots of text to explain what’s going on, it’s very clear and each potentially influencing variable you give it is assigned a percentage and is ordered by importance.
Whereas a data scientist would know exactly what data to feed the model in order to get the most accurate results, Power BI helps business users with minimal analytics experience. As an example, if you don’t give it enough observations for the chosen metric, the Key Influencer Visual will tell you that it needs more. You simply can’t do this kind of analysis with Microsoft Excel alone.
The Key Influencer Visual would be fantastic for a growing business currently using Excel or something similar – they could implement Power BI which is really cost effective. It’s exactly the kind of reason people should start looking to move from Excel to Power BI for their analytics and reporting because you’ll get so much more out of it.
Why Is This Feature So Groundbreaking?
Business users only get this level of analytics when they get into predictive analytics. The Key Influencers chart, in a very cost-effective way, gives them the ability to do some very advanced things – which is why everyone is absolutely raving about it.
The big thing for me is that without Key Influencer Visuals in Power BI, smaller businesses might have taken a lot longer to start using or benefit from this advanced kind of AI-driven analytics. But, because it’s a feature of Power BI which is cost-effective, accessible and available, it’s easy to get going with it and to start seeing the benefits of it very quickly.
What Other Benefits Does This Feature Offer?
You can build analytics based on the results of the key influencers chart. For example, if you didn’t realise location had such a big influence on customer churn, you could then create a report to proactively track that factor. Insights like these open management’s eyes.
Or, looking at your customer base and trying to predict which ones are likely to churn, the predictor variables you would enter into the predictive model would be the variables returned by the key influencers visual as being the most influential. Around 80% of predictive analytics is the data investigation and preparation, which could take hours of time, and the Key Influencers Visual simplifies this process. It helps with a part of that process which will ultimately free up data scientists to do more of the actual predictions.