This week’s list of data news highlights covers June 21-27 and includes articles about a computer vision algorithm that can detect rare genetic diseases in children and a Code for America effort to reduce emergency room calls in California.
Google launched a developer hub for its Nest smart thermostat hardware this week, including a public application programming interface (API) that will let third party developers to build apps using Nest data. A few companies already have products that work with Nest. Wearable technology company Jawbone offers a smart wristband that can tell when wearers wake up and signal Nest to set a temperature, and some Mercedes cars can connect with Nest to set temperatures when a driver is leaving or returning home. Developers are also working to use Nest’s API to design applications that control other household devices, such as Wi-Fi enabled light bulbs.
This week, the National Aeronautics and Space Administration (NASA) launched a competition to create new uses for the agency’s massive store of earth sciences data. NASA has released the data openly and offered computer resources to challenge participants through Amazon Web Services. The challenge, called OpenNEX after the NASA Earth Exchange scientific data collaboration platform, will offer $60,000 in prize money.
This week, educational technology leaders testified before a joint congressional hearing on student data. Mark MacCarthy of the Software and Information Industry Association and Thomas Murray of the Alliance for Excellent Education spoke to the House Education and Workforce Subcommittee on Early Childhood, Elementary and Secondary Education and the Homeland Security Subcommittee on Cybersecurity, Infrastructure Protection and Security Technologies. MacCarthy argued that new federal legislation around student data privacy was unnecessary given current laws and industry practices. Murray stressed the importance of education technology in preparing students for the workforce and discussed his own experiences as an educator using data in the classroom.
Fujitsu, a Tokyo-based company chiefly known for its computing and microelectronics products, has begun experimenting with smart farming. The company’s sensor-driven hydroponic lettuce nursery now produces between 2,500 and 3,000 heads of lettuce a day. The sensors monitor temperature, moisture, carbon dioxide content, air current, soil pH, and other indicators. By selling the finely tuned veggies at a luxury rate, the company hopes to make about $4.5 million in annual lettuce revenue by 2016.
Researchers at Carnegie Mellon University have developed an algorithm called LiveLight that can summarize long videos by extracting only the most interesting parts. The algorithm works by determining which moments in a video are visually novel compared to the rest of the video. Potential applications of the technology include summarizing surveillance camera video for easier monitoring. The algorithm’s creators have founded a startup to commercialize their method and speed up the algorithm’s complex data processing component.
The Obama administration announced this week that the Fraud Prevention System at the Centers for Medicare and Medicaid Services (CMS) recovered or prevented more than $210 million in improper payments in 2013. The system, which uses predictive analytics to scan through billing data and identify outlying cases and unusual patterns, also prompted CMS to investigate 938 health care providers and Medicare suppliers. Despite the success, the system has a long way to go: some estimates figure that improper Medicare payments amount to $50 billion annually.
Three Code for America fellows working in Long Beach, California, are using data to reduce the number of repeat emergency room (ER) callers. To identify these repeat callers, the team aggregated ER call data from the fire department and police department. They also added business license data to determine what kind of intervention might be best: the team reasoned that homes with lots of ER calls should be treated differently from restaurants, for example. The team is now visualizing the data to identify trends and generate actionable insights for city inspectors and health officials.
Computer scientists participating in a White House tech challenge want to equip disaster response dogs with wireless sensor equipment to facilitate communication with human emergency personnel. The sensors, which measure indicators including aerial gas content and the dogs’ heart rates, will allow first responders working with search and rescue dogs to learn more about what the canine helpers find in dangerous areas. The system is part of a larger emergency response communications platform that links drones aircraft, robots, dogs, and human first responders.
Researchers from Oxford University in the UK have developed an algorithm that analyzes photographs of children’s faces to detect whether or not they have rare genetic disorders. The algorithm looks for telltale facial structures and returns potential disease matches, which doctors could use to inform diagnoses. The researchers also used the algorithm to scan through a database of facial images and group patients together who are likely to have the same condition, even when no known diagnosis exists. This could potentially be useful in identifying ultra-rare genetic disorders.
The Food and Drug Administration (FDA) published a policy memo this month indicating that it will not regulate data systems in medical devices, deeming them low-risk to patients. The guidance does not apply to certain data-driven systems, including those that control delivery of medicine or oxygen to patients. Instead, the policy is targeted towards systems that store, transfer, or convert medical device data. FDA officials have expressed hopes that the move will encourage device makers to innovate without concerns about compliance.