Making It Easy For Data Scientists to Pinpoint and Share Visual Data Insight
Data Science is more important to business leaders than ever before. In fact, it’s now commonly accepted that data should be at the heart of almost every business decision. Still, business and customer data are only as valuable as the insight you can glean from it, and those insights are only useful when they can be communicated, shared, and acted upon.
That’s why we created a preview technology called #DataUnbound, which we presented to Adobe Summit attendees as a sneak peek back in March. “We need better ways to improve the accessibility of analytics,” said Sana Malik, a research scientist at Adobe. “Our team wanted to make it easier for data scientists to highlight the critical insight from their work and share it broadly.”
#DataUnbound is actually a collection of several different technologies that, when combined, makes it easy to identify key insights from data visualizations, automatically generate intelligent captions for those charts and graphs, and then share it with a few clicks.
It turns a process that used to take hours into one that can be completed in just a few seconds.
First, #DataUnbound uses Adobe Sensei, our artificial intelligence (AI) and machine-learning technology, to analyze charts created in Adobe Analytics, and generate intelligent captions that highlight the nuance of each data insight.
“It actually started as a project to improve the accessibility of analytics,” said Sana. “We wanted to see if we could create accurate, text-based descriptions of visual data.”
During her presentation at Adobe Summit, Sana showed how #DataUnbound can be used to analyze a line chart of exits from a website to online stores, not only accurately summarizing the data but uncovering a cyclical pattern that wasn’t obvious at first glance. “The AI analysis is applied to the chart itself, rather than the underlying data, and that’s what makes it unique — we’re able to capture the domain knowledge of the data scientist who created the visual in the first place,” she says.
The second key technology that goes into #DataUnbound is called Open Data Links. It allows the raw data underlying each chart to be easily exported into other data applications like Excel, Tableau, Power BI, or Terminal so that even ever-changing data stays current. “Data scientists need to be able to share their insight across all of their communication tools, so Open Data Link is critical for bringing the workflow together with all the latest updates and information,” says Sana.
Finally, the team created a series of plug-ins that allow the data captured by Open Data Links to be incorporated into tools like PowerPoint or Adobe Spark — with just a click of a button. That means data scientists can spend more time analyzing data, and less time creating presentation graphics or updating online dashboards.
As part of the Sneaks Presentations at Adobe Summit, Sana showed how #DataUnbound uses Open Data Links and a custom plug-in to generate theme-specific data visualizations with PowerPoint.
Although #DataUnbound is currently just a proof-of-concept technology, Sana says the team is already working to see how they can extend the technology to work with other types of charts and data visualizations. “There are so many other possible applications — from making the data inside PDFs in Adobe Document Cloud more accessible, to improving the workflow in Adobe Creative Cloud where you might want to create visual assets with up-to-date, data-based summaries,” she says.
#DataUnbound is the result of multiple collaborators, including Sana Malik, Adobe research scientist; Sungchul Kim, Adobe data scientist; Eunyee Koh, Adobe research scientist; and Courtney Ferguson, Adobe senior software engineer.
This story is part of a series that will give you a closer look at the people and technology that were showcased as part of Adobe Sneaks. Read other Peek Behind the Sneaks stories here.