Artificial Intelligence has been making its presence in various domains. The value derived out of AI has been making a positively disruptive impact in many industries. AI is now slowly getting embedded into self-service BI tools. With no exception to this trend, PowerBI is one of the early adopters for AI and Machine Learning Capabilities. We will be compiling a series of blogs covering the various AI options available in PowerBI.
In this blog, we’ll be discussing some of AI-powered visuals and capabilities which are available for the users off the shelf.
Key influencers is a built-in AI visual in PowerBI. Key Influencers help to identify the factors driving the KPI. For example, if we need to know, what influences my orders to increase or decrease, we could use this visual. This vision could be used to analyze hidden patterns in the data, which could be influencing your business decisions. It also features another tab named Top Segments which is used to see how multiple factors would influence the value. In the below example, orders are likely to decrease for some product categories and we can focus on categories in which we get more orders. Detailed documentation of this visual is available in the Microsoft blogs.
Decomposition Tree enables us to visualize data across multiple dimensions. It will automatically aggregate your data based on the dimensions enabling us to drill down into dimensions in any order. Microsoft brought in AI functionality to this by allowing users to drill down into the measure based on the highest or lowest contributor in the aggregation. This visual would handy for Ad-Hoc exploration or Root Cause Analysis. Detailed documentation of this visual is available in the Microsoft blogs.
Distribution Change in Bar Chart
Distribution Change is an additional feature available in the Bar Chart visual of PowerBI. It is useful in analyzing the distribution for a subset of the same dataset. You could check how division affects the distribution of cost per region. To use this feature, you will just have to right-click on the Bar Chart and click analyze and you can find how each category affects distribution. This will be helpful in identifying outliers from the general population. The information generated with this feature can also be added as a separate visual. Detailed documentation of this visual is available in the Microsoft blogs.
Q&A is an advanced visual in PowerBI which has the capability of understanding natural language questions and responding to them with appropriate visuals based on the result set. Like, the result set is of Date type then it would use the Line Chart to explain the result. Now Microsoft has given Q&A the ability to understand the context and understand follow-up questions like:
Q1. Who manufactures Tiguan?
Q2. Create a Bar Chart with City By Units Sold
Q3. How many customers bought the Yaris in Los Angeles?
Q4. In New York?
Q5. What about Tundra?
Q6. Show Houston instead
Q & A
We can either pin the resulting visual or keep the Q&A as a visual. We can train the visual to understand the questions with the help of Q&A tooling, also we can teach the visual the terms company personnel use. In addition to this, it comes with Row Level Security features. Detailed documentation of this visual is available in the Microsoft blogs.