Published on August 22nd, 2014 | by Travis Korte0
10 Bits: The Data News Hot List
This week’s list of data news highlights covers August 16-22 and includes articles about a Department of Transportation report on vehicle-to-vehicle communications and a UK effort to fund an Internet of Things communications standard.
The U.S. Department of Transportation’s National Highway Traffic Safety Administration (NHTSA) released a research report this week on vehicle-to-vehicle (V2V) technologies, systems which are designed to allow vehicles to communicate in real time. NHTSA says two potential applications stand out for early-stage testing: left turn assistance and intersection movement assistance, which would alert drivers to dangerous conditions in their respective circumstances. Those two V2V applications alone could prevent between 25,000 and 592,000 crashes and save between 50 and 1,083 lives per year. The report accompanied a notice of proposed rulemaking, which draws attention to policy and technical issues standing in the way of V2V adoption, including whether V2V systems should allow users to get information on mobile devices and whether V2V should solely be a component of the vehicle.
There is currently no universal standard that allows Internet of Things (IoT) devices to communicate and interoperate, meaning manufacturers have to develop ad-hoc solutions for connecting different kinds of devices. The Hyper/CAT standard, designed to solve this problem, is supported by a consortium of over 40 British-based organizations, including Rolls Royce, the BT Group, and the Open Data Institute. The UK’s Technology Strategy Board, a federal innovation agency that reports to the Department of Business, Innovation, and Skills, announced £1.6 million in new investment in the standard this week. The latest funding adds to £6.4 million the UK has previously committed to IoT standards efforts.
Rent the Runway is a New York-based startup that buys designer dresses and apparel wholesale and rents them inexpensively to customers for short-term use. The company uses data science to ensure its operations run efficiently, algorithmically tracking and forecasting demand for its 65,000 dresses and 25,000 accessories as they circulate among the company’s 5 million members. Rent the Runway’s founders see opportunities to expand their approach to different kinds of goods, hoping to one day become “the Amazon of rental,” according to the CEO.
The U.S. Food and Drug Administration (FDA) has released an application programming interface (API) to give developers access to its database of medical device incident reports. The database, which consists of thousands of reports dating back to the early 1990s, is already publicly available for download, but the API makes it easier to extract subsets of the data that are relevant to a particular analysis. The API comes with a visual interface where users can input a query, such as incidents for devices that include the term “x-ray,” and graph its results over time.
San Francisco-based startup Planet Labs makes and launches small, inexpensive earth imaging satellites. The company has launched 71 satellites in the past 16 months, and hopes to eventually have enough devices in orbit to image every inch of the Earths’ surface every 24 hours. Once it has this capability, the company plans to create a database that can be used to track environmental changes in near-real time, to which it will sell access to conservation groups, insurance companies, and other satellite data users.
The Defense Advanced Research Projects Agency (DARPA) has issued a crowdsourced challenge to develop computer models that can accurately forecast the spread of an infectious disease called Chikungunya. The disease, which was introduced to the United States this summer, spreads from person to person by mosquitoes, so the agency hopes challenge participants can use network science and other epidemiological techniques to model its proliferation. Winning teams can earn up to $150,000. The challenge organizers are explicitly seeking solutions from outside the public health community, in hopes of finding modeling approaches that have been useful in other applied data science fields.
Berkeley-based startup Chemisense is building a wearable air-quality monitor that uses connected sensors to report airborne chemical information to a smartphone app. Users who live in highly polluted areas or have respiratory conditions can check the app to keep track of their exposure to dangerous chemicals. The company’s first target market is parents of children with asthma, who Chemisense hopes will want to identify which areas in their neighborhood provoke attacks. The company expects the first prototypes to be manufactured in the next six months.
Reddit, along with image hosting site Imgur and video game streaming platform Twitch, have formed the Digital Ecologies Research Partnership, which will provide researchers with user behavior data from those sites. The partnership’s organizers hope to help promote research into online social dynamics and have stated that all research conducted under the partnership will be released openly. The organizers will also offer their data to researchers free of charge.
Hampton Creek, a San Francisco-based food and technology startup, plans to build the world’s largest database of plant information in hopes of encouraging food producers to create vegan replacements for foods that use animal products. Dan Zigmond, the former lead data scientist for Google Maps, will lead the company’s data efforts. The company will analyze the structure of plants’ chemical bonds, their reactivity with water, and other properties to determine their suitability for certain products, such as mayonnaise or cookie dough.
Information designer and noted “quantified self” experimenter Nicholas Felton released a data-driven report cataloging his communications and other activities over the past year. Felton’s data set consisted of 94,842 interactions that he logged, including telephone calls, emails, in-person conversations, and even nonverbal communication. Felton’s findings included the fact that even though his email and text message quantities were comparable, he interacted with far more email addresses than phone numbers.
Photo: Flickr user Matthew Rutledge