10 Bits: The Data News Hot List
This week’s list of data news highlights covers March 15-21 and includes articles on Facebook’s research into facial recognition software and a UK government-funded data science institute.
This week, the White House launched its Climate Data Initiative, an effort to draw on open government data and develop tools to help local communities anticipate, prepare for, and respond to extreme weather events. The White House also launched climate.data.gov, a website which will consolidate data from the National Oceanic and Atmospheric Administration, the National Aeronautics and Space Administration, the U.S. Geological Survey, and other sources. Private partners will play a major role in the initiative. For example, mapping company Esri has committing to work with 12 cities across the country to create open maps and apps, and Google will donate one petabyte of cloud storage for climate data.
Facebook released a research paper earlier this month detailing a facial recognition software product called DeepFace, which can determine whether two images show the same person with over 97 percent accuracy. The software, which represents a considerable improvement over previous face-matching software, uses deep learning algorithms that help data scientists automatically construct complex models to make sense of large datasets. The software was developed at Facebook’s artificial intelligence lab, which launched in December 2013.
This week, the UK government announced the creation of a data science research institute, along with a £42 million ($69 million) investment over five years. Government ministers hope the institute, which will bear the name of pioneering English computer scientist Alan Turing, will cement the UK’s place in the global data science community and drive economic growth in the nation.
A gas explosion in New York City’s East Harlem neighborhood earlier this month claimed the lives of eight people and left at least 70 others injured. Within seconds, Twitter was able to confirm the incident based on the rapid increase in tweets about an explosion in that area. Although individual tweets are not always trustworthy, aggregate tweet information can be highly reliable. Twitter and other companies have developed algorithms to automatically identify tweets around similar breaking news topics and rapidly flag them as newsworthy.
The U.S. Geological Survey will soon combine two of its major mapping programs into a single source for geospatial and cartographic information. The National Atlas, which offers map data covering agricultural, biological, geological, social and other variables, will be absorbed into the National Map, which acts as the definitive store of federal topological data. The National Atlas will be removed from service on September 30, 2014. Officials hope the consolidation will provide simpler access to the two key data sources and cut costs for the agency.
Researchers have used Facebook data to track the spread of fictitious news stories, finding that people who engage with content on alternative news sites are much more likely to participate in conversations about false stories. The research looked at anonymized data from over 1 million people during the Italian elections of 2013. The researchers identified specific news sources known to post satirical or otherwise false content, and tracked users’ interactions with those posts. The researchers found their results surprising, as consumers of alternative news sources often cite a fear of misinformation from mainstream news sources as one reason they read alternative sources.
This week the United Nations Food and Agriculture Organization (FAO) launched a new database that consolidates various sources of data about land cover, including information on whether a particular region is covered by agriculture, trees, forests, or bare soil. The Global Land Cover SHARE database draws from countries that collected the data with different standards and in different formats, and has reconciled these sources into a single, authoritative source for land cover data. UN officials say that this database is crucial to promoting sustainable resource management across the globe.
A technology for measuring voice modulations to detect deception may supplement polygraph tests in the future. Computer voice stress analysis, which tracks changes in subjects’ speaking patterns that may indicate stress, has been developed for applications in border security as well as monitoring fatigue in airline pilots. One such system, which was piloted by Romanian border control authorities in 2013, also collects data on eye and body movements to create a fuller picture of emotional and physical stress.
A U.S. interagency report released this week outlines a plan to connect computer systems that are used to issue security clearances and identification cards. The plan, prompted by the Washington, D.C. Navy Yard shootings that killed 12 people in September, 2013, includes real-time alerts when problems are detected and strives to create an environment in which security checks are continuous rather than periodic. The plan will also integrate additional data sources, including social media, to keep abreast of events that may have security implications.
The SpotHero app helps Chicago drivers navigate the city’s complex parking environment. The app, which lists parking lots, meters and even private spaces that owners rent out, also lets users compare prices and reserve certain spots remotely. The app won the top prize in a recent app contest, and it has begun spreading to seven more cities in the Midwest and Mid-Atlantic regions.