There is an abundance of hype around artificial intelligence (AI) and machine learning (ML) in and around the analytics market. Matt Aslett and Krishna Roy published a new 451 research report last week highlighting a few key areas where AI is “assisting” humans for decision making (rather than just making the decision for them). “AI for BI: A potentially harmonious marriage, but not yet problem-free” is a nice snapshot and covers three key areas: “smart” visualizations, approaches for weaving AI into storytelling (incorporating NLG and NLP), and finding unknown insights and alerting users.
You can think of AI for analytics like “GPS navigation” for business users who have to deal with tons of information and big data, and need some help navigating to find useful insights. This leaves the power of decision making (i.e., “driving”) in the hands of humans, but it makes them faster and smarter in their quest for knowledge and faster decision making.
I thought it might be helpful in this blog to double-click into what 451 calls “smart visualizations” so you can see it in action. We implemented this in our latest software release of Arcadia Enterprise; I use it myself in my own job and find it incredibly helpful and cool.
First, what the heck is “smart visualization” and why is it helpful? 451 points out,
“…until fairly recently certain visual analysis offerings still required the user to know the most appropriate visualization to employ to most clearly depict the answer to their question. While that was fine if the end user was a data analyst – or someone else with solid data and analysis skills – those that weren’t analytically literate were often left scratching their heads and guessing the right visualization to use. Through trial and error, they could be successful and produce an infographic that was not just pretty but pertinent and useful. However, if they failed, misconceptions about the insights within the visualization could occur.”
Yep. I’ve spent weeks trying out different ways to visualize data. Wouldn’t it be great to see YOUR data laid out with multiple visualization options in one view, like a palette of choices which you can select and it instantly applies? And, even better, wouldn’t it be great if they were recommended based on visualization best practices and took into account cardinality and other aspects in the dimensions of the data that you were trying to visualize? Well, that’s what “smart visualizations” are.
As 451 describes it, “The introduction of machine learning into the visual analysis process to produce ‘smart’ visualizations has emerged as a key way to apply AI to BI to eliminate this knowledge gap. A smart visualization is so called because it is designed to provide a best-fit data visualization so the user doesn’t have to work it out. Smart visualizations essentially add AI to the process of selecting the right graphic for the right query via machine-learning-based suggestions. They are becoming a key way to enable personnel without deep analytical skills to create dashboards with good infographics more quickly and easily.”
In Arcadia Enterprise (or our free desktop tool, Arcadia Instant), we call these “Instant Visuals.” Our goal is that users spend less time iterating on which visual type works best and can spend more time analyzing and optimizing their business. See it in action! Here’s a short 2 minute video showing how you can quickly analyze billions of rows of television viewing data to come up with interesting insights.
Video: Instant Visuals – AI-driven visualization recommendations for a given data set
You can give it a try for yourself and download Arcadia Instant for free, unlimited use on your desktop (Mac or Windows). We’ve tried to make it even easier to get started with new tutorials and knowledge base articles (such as Four Tips and Tricks to Remember for Arcadia Instant).
The future of AI for analytics is here – have fun!