10 Bits: the Data News Hotlist
This week’s list of data news highlights covers December 12-18, 2015 and includes articles about new regulations in California that could restrict the development of self-driving cars and how U.S. states are turning to data technologies to improve snow removal efforts.
The U.S. Food and Drug Administration (FDA) has launched precisionFDA, a cloud-based web portal to make it easier for researchers throughout the United States to collaborate on developing next-generation sequencing (NGS), a genetic analysis technique that could lead to more personalized approaches in medicine. Users of precisionFDA will be able to access a variety of useful data, such as reference genomes that serve as a baseline for comparing genetic data, and software tools to analyze genetic information. In addition, they can use the portal to share their results and solicit feedback. The FDA will also make the code that runs precisionFDA open source via code repository website GitHub to encourage the public to improve the portal.
Netflix is developing a new streaming algorithm to reduce the amount of bandwidth required to stream videos without sacrificing video quality to make its streaming service more efficient and better serve customers with data caps on their Internet usage. While Netflix already automatically adjusts its algorithm to accommodate users with slower Internet speeds unable to stream high-definition videos, its new approach will take into account that different types of content have substantially different data needs. For example, animated shows typically have much less visual detail compared to an action movie that has lots of explosions, which require more data to display. By streaming content with Netflix’s new algorithm, customers could watch the same content while consuming 20 percent less bandwidth.
NBCUniversal has partnered with TiVo and market research startup RealityMine to collect and analyze data about how audiences watch the 2016 Summer Olympics across different platforms, such as television and mobile phones. The partnership will study how new digital platforms influence viewership habits, such as how people choose to watch the Olympics outside of their homes and how social media influences Olympics coverage. The researchers involved in the partnership expect studying the Olympics will provide insight into the future of how new mobile technologies will shape viewership trends.
In July 2014, the Indian government announced an ambitious plan to develop 100 smart cities. As of the December 15 deadline, India’s Ministry of Urban Development had received 85 smart city proposals from states and union territories. The government has allocated $7.4 billion over the next five years to finance the development of these smart cities, and it has called on local governments to match national funding.
The California Department of Motor Vehicles (DMV) has proposed new regulations that would prohibit self-driving cars without a brake pedal or steering wheel, as well as make it illegal for them to operate without a licensed human driver present and able to take control of the car. The DMV justified the proposed rules by claiming the technology is too risky for use by the general public at this stage, despite claims from Google, which manufactures self-driving cars in California, that their autonomous vehicles are safe and that such rules would limit the potential of the technology. Additionally, the proposed rules would require manufacturers of self-driving cars to report to drivers the type of data they collect from the vehicle.
Booz Allen Hamilton and data science competition company Kaggle have launched the second annual Data Science Bowl, an international competition to develop new data tools to solve pressing challenges, this year focusing on improving cardiovascular disease diagnostic techniques. In partnership with the National Institutes of Health, the Data Science Bowl will challenge participants in the 90-day competition to develop an algorithm that can automate measurements in magnetic resonance imaging (MRI) scans that can indicate heart disease. The participants that develop the best algorithm by the end of the competition on March 14, 2016, will receive a $125,000 prize, and runners up will receive an additional $75,000 in prizes.
Chinese Internet-search company Baidu has developed Deep Speech 2, a deep learning system capable of recognizing English and Mandarin speech better, in certain conditions, than humans. Baidu trained Deep Speech 2 with thousands of hours of transcribed audio recordings too make it capable of translating speech and even recognizing new words it has never heard before. Deep Speech 2 will help Baidu improve its speech-recognition apps, which are particularly popular in China because text entry can be time consuming and some users are unfamiliar with Pinyin, the method for transcribing Mandarin using Latin characters based on phonetics.
The International Telecommunication Union (ITU), a United Nations agency that monitors information technology, and the Georgia Institute of Technology (Georgia Tech) have partnered to spur the development of interoperable standards for applications of the Internet of Things useful in solving urban-development challenges around the world. ITU and Georgia Tech will bring together standards groups, trade associations, and public and private sector groups working with the Internet of Things to foster discussion about how to best establish interoperability across different industries and markets.
Medical institutions in the United States are required to report the results of clinical trials to ClinicialTrials.gov within a certain time period so that the medical community can use this data to evalue the safety and merits of various treatments. However, an investigation by STAT, a news publisher focusing on health and medicine, revealed that many prominent institutions, such as Stanford University and the University of Pennsylvania, failed to disclose this data, or disclosed it past the deadline, 95 percent of the time. Additionally, STAT found that these violations could warrant a total of $25 billion in fines, but the federal government has yet to levy a single fine against an offending institution.
As U.S. state transportation agencies prepare for the winter, many are turning to data technologies for the first time to improve snow clearing efforts by reducing costs, protecting the environment, and making roads safer. Several state agencies have developed weather stations that collected data from sensors embedded in pavement to monitor road temperature, dryness, and whether or not it has already been treated with snow-melting chemicals to help officials make more informed decisions about snow-clearing efforts. States such as Pennsylvania have also equipped hundreds of snow-plowing trucks with sensors that can track their location and monitor road conditions to better manage their fleets. And the Minnesota Department of Transportation has developed an algorithm to recommend the most efficient combination of snow-melting chemicals and plowing frequency, which can reduce costs and chemical usage, based on data its snow plows collect on atmospheric conditions.