The U.S. Food and Drug Administration this week announced a plan to make its publicly available data sets accessible...
electronically for the first time, in order to unleash application and data visualization developers on the information.
The data will be released in structured, computer-readable formats through a new program called openFDA. In a press release, FDA officials said they hope developers will use the data to create mobile applications, data visualizations and consumer information campaigns that address adverse drug interactions and other public health issues.
"These reports will be available in their entirety so that software developers can build tools to help signal potential safety information, derive meaningful insights and get information to consumers and healthcare professionals in a timely manner," said Taha Kass-Hout, chief health informatics officer at the FDA.
The data sets will be released in stages, starting with a pilot project involving millions of reported incidents of adverse drug reactions and medication errors. Until now, the FDA said, that information was available to external users only in unwieldy reports or through Freedom of Information Act requests. Databases on product recalls and product labeling will be added to the pilot project at an unspecified later date.
Ultimately, the data dump will provide a complex big data puzzle for outside researchers to try to solve. FDA officials are hoping that by letting external users sift through the mountains of information, they'll be able to identify previously overlooked correlations that may be clinically relevant.
Brian Norris, CEO of Social Health Insights, which analyzes social media data for public health purposes, produced one of the first working applications for openFDA, a search tool that lets users look for reports and statistics on adverse events. He is documenting his findings on Twitter, where he uses the handle @Geek_Nurse:
I just clicked the "Show Me The Data" button and got analytics on nearly 2 Million Adverse Event Reports in < 1 sec! pic.twitter.com/Gk2PEB629N
— Brian Norris (@Geek_Nurse) June 2, 2014
IBM helps colleges tackle data scientist shortage
It's well known that businesses are having a difficult time finding enough workers with data skills to lead their big data and analytics initiatives. But some colleges, with help from IBM, are trying to address the big data skills shortage.
More on data analytics and management
Podcast: Should we stop talking about big data
Don't overlook the potential downside of big data
Interview with Curt Monash on real-time analytics and Spark
IBM announced late last month that it is partnering with 28 business schools and universities for the upcoming 2014 fall semester. IBM is giving the schools access to technology and helping them craft programs designed to give students the skills they will need to pursue careers in data science.
For example, IBM is working with Case Western Reserve University to develop an undergraduate program in data science and analytics aimed at developing industry-specific skills for working with big data. The company is also working with the University of Massachusetts, Boston, on a Master of Business Administration degree with a specialization in business intelligence.
Bob Picciano, IBM's senior vice president of information and analytics, said the move is all about training the workers that businesses say they need.
"Taking advantage of the transformation opportunity presented by big data and analytics has become a key priority for organizations," Picciano said. "To embrace this growing opportunity, companies today must hire a workforce with a broad range of big data and analytics expertise."
Our World in Data, a site devoted to cataloguing and visualizing how humans are changing the world, has a set of new visualizations looking at historical data on natural catastrophes. The collection of visualizations looks at the historical record of natural disasters, such as earthquakes, hurricanes and volcanic eruptions, and shows their correlations with deaths. Take a look at the visualizations here.