“My career for the past 25 years has been getting useful information from data through interactive visual representation. So, I’ve been handling big data for a very long time,” says David Ebert, Silicon Valley Professor of Electrical and Computer Engineering and director of Visual Analytics for Command, Control and Interoperability Environments, or VACCINE, a U.S. Department of Homeland Security center based at Purdue.
The work dovetails with a College of Engineering initiative called Data Mind, which focuses on enhancing existing efforts in research and education, as well as launching new programs in data science and its interface with engineering. The initiative is linked to the university-wide goal of establishing Purdue as a leader in the broadly construed field of data science.
“When you talk about data and mind, the whole goal of visual analytics and the research we do is to create a collaborative decision-making partnership between automated computerized algorithms and the humans who are using the system,” Ebert says. “We create interactive visual representations, turning data into something the human can understand, so the human and the computer are collaborating to make better decisions.”
Ebert’s group has helped people in many different areas harness data to solve problems more effectively — from law enforcement in several communities, to better allocating resources, to solving and reducing crime, to agricultural producers enhancing use of scarce water and labor resources, to health, environmental and weather researchers making new discoveries.
VACCINE’s research has also allowed the U.S. Coast Guard to improve situational awareness, resource allocation, budgeting, planning and response effectiveness. The Government Accounting Office has recently determined that a VACCINE developed search-and-rescue-optimization tool will save the Coast Guard an estimated $290 million over the next 10 years.
One of VACCINE’s latest creations is the Social Media Analytics and Reporting Toolkit, or SMART, designed to better manage a diversity of mayhem, including filtering out hoax calls during disasters, to dealing with “kayaktivist” protesters in the Pacific Northwest; and keeping the peace during large public events like the Republican National Convention; the Presidential Inauguration; and baseball’s All-Star game. SMART allows officials to more quickly respond to bomb threats, shootings, natural disasters, fires and various other catastrophes. The tool, for example, makes it possible to filter thousands of tweets and Instagram posts for information about natural disasters. It is allowing the Coast Guard to efficiently dispatch emergency personnel to citizens who are truly in need of rescue.
“If someone calls on a radio and says, ‘SOS, my boat is sinking, I am at this location,’ the Coast Guard needs to send people to respond. There has been a string over the past two years of people making fake SOS calls, and that means the Coast Guard is launching helicopters and boats, which costs tens of thousands of dollars and puts people at risk,” Ebert says.
SMART scrutinizes social media chatter looking for these hoaxes. “People may brag about having done this, or someone might have heard or seen something and posted about a hoax,” he says.
The tool was developed in work led by Ebert and a team of graduate students and postdoctoral research associates in VACCINE.
“SMART is gaining a lot of traction right now to monitor social media, and we are taking this tool from a prototype development stage into something that’s commercially viable and sellable to first responders,” says Ebert. “The goal is also to make SMART available for download by any first responder.”
The tool is enabling officials to discern the vast tangle of unrelated information from the nuggets of valuable, potentially life-saving postings.
One takeaway from the research is that social media may help authorities monitor disaster zones when public cameras and other sources of information are knocked out. The social media pictures and videos can crowd-source on-the-ground reporting and provide actionable information on conditions across a community, when properly filtered with our tools.
Data from public cameras also might be used in conjunction with social media to precisely determine the hardest-hit areas. Ebert’s team has partnered with ECE Professor Yung-Hsiang Lu’s team to link the SMART tool with his work in harnessing public video cameras.
Lu’s new system allows law enforcement and public safety agencies to tap into thousands of cameras accessible on the Internet and located in numerous venues, including parking garages, national parks and highways. In addition to applications in law enforcement, the system can be used to quickly find damage, plan rescues and conduct other operations during natural disasters, such as tornadoes, hurricanes and blizzards.
On the education side, Ebert is proposing freshman-level classes in visual data analytics to teach students how to interpret huge volumes of information to produce accurate and meaningful visualizations.
“As you become a true data scientist, you need to know how to interpret data in all forms,” he says. “It’s very easy to mislead people with visualization, and that becomes a problem with people using more tools with data. I don’t want a surgeon to operate on me based on the color of a pixel in a visualization. I want them to be able to go back and get the actual scanner’s value and make sure that it’s in a range that is a problem.”