This week’s list of data news highlights covers May 31-June 6 and includes articles about Apple’s foray into the Internet of Things and a Secret Service effort to distinguish sarcastic threats from real ones on social media.
MIT Media Lab’s StreetScore algorithm assigns Google Street View photos a score based on how safe they look to humans. Using crowdsourced data on safety perceptions of streets in cities around the country, the algorithm’s creators have launched a pilot website showing scores in New York City and Boston. The data the algorithm generates could be useful in informing urban design decisions in the future or exploring the correlation between urban perception and crime.
Edyn, a Berkeley, California-based Internet of Things startup, wants to build a smart garden sensor to help gardeners make planting and upkeep decisions in real time based on environmental data. The company has launched a Kickstarter campaign to fund its soil analysis and watering system, which can transmit its data to gardeners through Wi-Fi. In addition to helping gardeners raise healthier plants, the system could help them use less water. This could be a major boon to the water supply, since as much as half of the 7.8 billion gallons of water used annually for landscape irrigation in the United States goes to waste.
A new international standard known as ISO 37120 lays out precise definitions for 46 measures cities can use to compare their performance. The standard, which covers variables including airborne particulate content, life expectancy, and green area per 100,000 residents, will help ensure that municipal data collections in one city can be directly compared against information from other cities. Cities will not be legally required to adhere to these standards, but they will likely face pressure from business and academic groups who rely on data validity and comparability to make decisions.
The Secret Service is soliciting proposals to analyze mentions of the agency and the president on social media and determine which of the many mentions represent genuine threats. In particular, the proposed systems must be able to detect sarcasm, which is an active area of research in text analytics. In addition to detecting threats, the monitoring system could help the agency monitor public opinion if events such as the 2012 prostitution scandal recur in the future.
Doctors at the University of Arizona Medical Center have developed a wearable “smart sock” to help diabetes patients identify lower extremity infections before they lead to amputation. Diabetes can cause a loss of sensation in limbs, which means that such infections often go undetected. The device monitors pressure levels in patients’ feet to identify potentially dangerous ulcers and alert patients that they require medical attention. Researchers expect the devices to be available to patients by 2019.
Skymind, a San Francisco-based startup, launched this week to offers support and services around its open source deep learning tool. Deep learning, a branch of machine learning loosely inspired by the way information is transmitted in the human brain, has recently been a hot topic among large technology companies such as Facebook and Google, but currently requires considerable academic expertise to deploy. Skymind hopes its system, called deeplearning4j, will make deep learning more accessible to smaller companies.
The U.S. Food and Drug Administration (FDA) launched OpenFDA this week, a platform to provide access and search capabilities to the agency’s open data. OpenFDA offers developers an application programming interface (API) with which they can develop their own applications using the data. The FDA also released more than 3 million reports on adverse drug events and medication errors to coincide with the platform’s launch. These datasets were prioritized because they were found to have been the most frequently requested through Freedom of Information Act requests.
Banking and credit card company Capital One is using data analysis to inform its hiring process. The company has spent the last two years surveying every job candidate about their satisfaction with various aspects of the application process, hoping this approach will help it improve application experiences and attract top talent. One insight gleaned from the analysis is that candidates often had trouble recounting benefits details to their families. As a result, the company began offering take-home materials on those topics.
At its worldwide developer conference this week, Apple unveiled plans for a product called HomeKit, which attempts to reconcile the various security and communications protocols of the Internet of Things into a single interoperable system. This has historically been difficult owing to the wide variety of technologies used to control and communicate between connected devices. The company will offer the app to iPhone and iPad users, who will be able to get data from and send instructions to a variety of Internet-connected devices using no additional controllers.
This week, the Office of the National Coordinator for Health Information Technology (ONC), a division within the U.S. Department of Health and Human Services, released a 10-year plan for health information technology interoperability. By 2017, the agency expects to develop a roadmap to scale health data exchanges across different technology providers. By 2020, it anticipates that health care providers will be able to share health data quickly and easily, and integrate data from a wide variety of sources into a single record. By 2024, it hopes patients will be able to manage their own health data using personal devices and share that data with third parties as necessary.