Advanced Analytics in Power BI – AI in BI
Data has been a part of our lives for a very long time, and Business Intelligence (BI) has been a popular tool for helping organizations derive insights from it. These days, AI is reshaping how we use data and how organizations thrive. While BI serves as the backbone of decision-making, AI excels at discovering patterns and forecasting. A good combination of the two can fuel better decisions.
In this upcoming blog series, let’s delve into the concepts of Natural Language Processing (NLP) and its application within BI. There will be a couple of effective methods to integrate AI into your BI projects using Power BI, providing practical guidance and real-world examples. Furthermore, let’s explore some important considerations and challenges to be mindful of when leveraging AI in BI initiatives.
NLP and its business application
One component of AI is Natural Language Processing (NLP), which focuses on the interactions between human language, known as natural language, and the computer. The primary objective is to enable computers to comprehend and respond to natural language inputs by parsing through extensive datasets. These datasets can be in form of text and speech and more. By leveraging semantic structures embedded within the data, computers decipher the contextual relationships and meanings between words, and then close the gap between human language and machine understanding.
NLP is already in our life. You must have heard of or used Alexa, Siri, web search engine, etc. NLP also has many possible business applications:
Now that we understand what NLP is, let’s explore how we can leverage it when working on business intelligence projects. NLP serves as a transformative component, empowering businesses to engage with their data directly. With NLP, data consumers can extract insights from data through simple, human-like queries. NLP not only understands human language but is also able to interpret data and visuals. In the next section, I will guide you through the process of creating and interpreting NLP-powered visuals in Power BI.
Q&A visual example walk through
There has always been a desire within the business side of the organization to engage more effectively with data. In Power BI, there is a Q&A visual that enables users to ask questions in their own words directly to the data. Business users can engage with data more effectively, while dashboard developers can draw inspiration from its insights when visualizing data.
The sample visuals below are based on an airline data from Kaggle[1]. This dataset contains various parameters about airline operations, such as passenger ID, gender, age, nationality, departure airport continent, departure date, flight status, etc. Here is a screenshot of a couple of entries from the dataset.
After selecting the Q&A visual, a visual with a text box and a couple of suggested questions will pop up. If you find a suggested question interesting, you can simply click on it. You can also type in the question in the text box. In this example, I clicked on “What is the number of passengers by continent,” and the Q&A visual displayed a bar chart. Power BI will pick the visual type based on its understanding on the data.
If you have a specific visual type in mind, you can include the visual type in the text box. In this example, I added “pie chart” to the text box, and the Q&A visual returned a pie chart. If you are satisfied with the result and would like to include the visual in your dashboard, you can click on the “convert” button in the upper right corner. The visual will be converted to a standard visual type, in this case, a pie chart.
The Q&A visual also provides options to personalize your dashboard. By clicking on the gear icon in the upper right corner, you will access the “Q&A setup” window.
- Synonyms: You can add or remove terms that might be used as synonyms for tables and fields.
- Teach Q&A: Provide Power BI with definitions of questions and terms people may use.
- Review questions: This is a list of questions people have asked. You can use this information to understand what popular questions are and refine your metrics.
- Suggest questions: Add suggested questions that will show up on the first window of the Q&A visual.
Smart Narrative visual example walk through
BI analysts typically invest a significant amount of time in creating compelling plots and tooltips. However, there remains a risk of dashboard users misinterpreting the plots. The Smart Narrative visual serves as a text box that provides explanatory information on the dashboard page, such as trends, percentage change, maximum, etc. It will help to mitigate misunderstandings and enhance comprehension. Developers have the flexibility to edit and format the text to suit their audience.
Below is a sample dashboard created using the airline dataset. In the upper left corner, you’ll find two slicers – one for departure date and another for departure continent. Additionally, there are two plots: a bar chart breaking down the number of passengers by their departure continent, and a line chart displaying the number of passengers over departure date. In the upper right corner, there is a textbox-like visual, which is a Smart Narrative visual. The Smart Narrative visual automatically analyzes all the visuals on the page and summarizes highlights of all the information.
To make it easier to see, I have highlighted the text and where the insight was derived in the screenshot below. You can see that the text highlighted in green is derived from the line chart below. There is a downward trend on the number of passengers over their flight departure date. The summary in the blue box is from the bar chart – Continent North America has the greatest number of passengers.
Let’s try playing with the slicer on the left – select a shorter time frame and a couple of specific continents. You can see that the text box will be updated according to your selection on the slicer. Now the overall trend in the selected timeframe and continents is upward, and Asian has the most passengers.
AI is an incredibly powerful tool that can significantly enhance our ability to understand data and establish a robust foundation for reporting. As discussed in this blog, integrating AI with BI in Power BI can revolutionize how businesses interact with and derive insights from their data. Stay tuned for our next blog in the AI in BI series, where we will delve deeper into using AI for diagnostic and predictive analytics in Power BI. In the meantime, discover how ProCogia’s Data Analytics services can transform your Business Intelligence efforts and help you stay ahead of the curve.
