This week’s list of data news highlights covers May 10-16 and includes articles about a venture capital firm that appointed an artificial intelligence system to its board and a startup that hopes to reduce traffic in cities around the world.
1. Using Medicare Data to Help Vulnerable Patients During Disasters
The Centers for Medicare and Medicaid Services has launched a program in which federal officials can access health data to identify and work with cities to assist vulnerable patients during emergencies. The program was piloted during an ice storm that threatened New Orleans this January. During that event, kidney dialysis patients and local residents who relied on breathing machines were identified and advised on treatment options and how to find help during the storm. The Assistant Health Secretary for Preparedness and Response plans to scale the program across the country. In the future, the Department of Health and Human Services also plans to release in interactive map indicating how many Medicare beneficiaries have wheelchairs and other equipment in various areas, in order to help health officials plan where to build shelters and prioritize resources.
2. Startup Uses Data to Reduce Traffic in Cities Around the Globe
Bay Area startup Urban Engines thinks it can use data to help cities reduce traffic. The company, which launched publicly this week, is backed by some of Silicon Valley’s largest venture firms and has held pilot studies in Bangalore, Singapore, and Palo Alto. Initial experiments used transit fare card data to develop a real-time picture of commuter activity; future efforts will integrate vehicle-based information as well.
3. MIT Researchers Develop Algorithm to Automatically Understand Video
MIT researchers have developed an algorithm that can identify what is happening in videos. The researchers’ approach takes the video frame-by-frame, identifies objects, and attempts to guess what is going on from the combination of those sets of objects in a particular order. Computer vision, particularly with video, has traditionally been a difficult problem in computer science, but the MIT algorithm joins a number of other recent approaches that could eventually be applied in contexts from medicine to law enforcement.
4. IBM Watson May be Coming to a Dermatologist’s Office Near You
IBM announced this week that it will license its Watson artificial intelligence platform for use as a diagnostic tool for dermatologists. Watson, which is best known for defeating two of the most prominent human Jeopardy players, has begun to find applications in health care decision support systems. IBM is partnering with Modernizing Medicine, a Florida-based electronic health record app company, to integrate Watson into their dermatology software. Watson will accelerate note-taking and data collection, as well as offer automatically generated diagnosis suggestions based on what it learns from doctors’ descriptions of patients.
5. Venture Firm Appoints Artificial Intelligence System to Board of Directors
Hong Kong-based venture capital firm Deep Knowledge Ventures has appointed a machine learning software program to its corporate board. The program, called the Validating Investment Tool for Advancing Life Sciences (VITAL), analyzes financing trends in life sciences companies to help the venture firm, which specializes in those areas, make smarter investments. The company has already used VITAL to inform its investments in two companies. The ultimate goal, according to the third-party provider that created the software behind VITAL, is to refine the program until it can autonomously manage an investment portfolio.
6. Food Nutrient Sensor Hits Big on Kickstarter
A tiny sensor that gathers data on food went viral on crowdfunding platform Kickstarter this week, amassing over $1.5 million dollars in support for a proposal that only sought $200,000. The device, called SCiO, was developed by Tel Aviv-based company Consumer Physics and promises to detect calories, fat, carbohydrates, protein, and other food contents. The creators hope it will also be applied to verifying pills’ contents. Although it does not yet support all foods and medicines, the creators encourage early adopters to add new entries to its already sizable database and continually improve the product.
7. Wearable Technologies for Infant Health
Wearable technologies that monitor infant’s vital signs are coming on the market. Owlet Baby Care’s “smart sock” and Rest Devices’ Mimo Baby bodysuits measure indicators such as respiration, skin temperature, oxygen levels, and heart rate. The devices then send the data to smart phone apps, where parents can monitor their children’s health remotely. Both companies are wary of the tight regulations on children’s products, adding mechanisms to detect when their devices overheat and monitoring battery energy usage.
8. Data from Doctor Visits Driving Diagnosis Research
Researchers are using data collected from routine checkups to improve doctors’ diagnoses. In one instance, researchers used data on patients with sore throats to develop a method that recommends either a strep throat test or an ordinary aspirin pill. The data, which is anonymized, also lends itself to applications from optometry to immunology. The increasing adoption of electronic health records among doctors and hospitals has enabled these applications, making it easier for researchers to analyze large amounts of standardized data that had previously been stored on paper or in other less useful formats.
9. Emerging Education Programs in Data Science
Eight new data science education programs have launched or will be starting up in 2014. These include both masters degrees and certificates, and span online and classroom formats. The programs span a variety of focuses, from public policy to “big data”-specific technologies. The universities offering the programs are Indiana University, UC Berkeley, Saint Peters University, George Mason University, the University of Chicago, the University of Southern California, the University of Virginia, and Worcester Polytechnic Institute.
10. Making Roads Safer for Bikes with Data
Two University of Virginia graduate students undertook a research project analyzing large quantities of traffic camera data that they hope to improve bicycle safety. Using computer vision techniques, they trained a computer program to recognize bicycles from cars in camera footage. The students incorporated this information into traffic models they hope will one day help city planners build roads that are safer for bicyclists. The project is one of the inaugural efforts funded through the university’s recently founded Data Science Institute.
Photo: Heb / Wikimedia Commons