This week’s list of data news highlights covers October 24-30, 2015 and includes articles about how the United States plans to use data to protect against dangerous solar flares and how machine learning could prevent gambling addiction.
The White House has published its third Open Government National Action Plan to demonstrate how the United States is fulfilling its commitments to the Open Government Partnership, an international initiative started in 2011 to promote transparency and encourage the use of technology to improve governance. The plan outlines the Obama administration’s existing efforts to expand the use of open government data and describes new projects the administration will undertake, including publishing information from the tax forms of nonprofit and charitable organizations in open and machine-readable formats.
The U.S. National Oceanic and Atmospheric Organization (NOAA) has partnered with Amazon Web Services to publish its entire collection of Next-Generation Weather Radar (NEXRAD) data, dating back to June 1991, online for anyone to use. The NEXRAD data, rated by the National Science and Technology Council as the government’s second most valuable geospatial data set, consists of 270 terabytes of geospatial information about precipitation and atmospheric movement. NOAA published this data with Amazon Web Services as part of NOAA’s Big Data Project, which enables NOAA to publish large data sets for public use with the help of private sector cloud service providers.
Gaming analytics company BetBuddy has partnered with researchers from City University London to develop a machine learning system that can warn online gamblers if they demonstrate signs of addiction. The software analyzes data about user habits from online gambling websites and compares it with data from users that have asked to be blocked from these sites out of concerns that they had become addicted. Innovate UK, the United Kingdom’s government agency devoted to fostering innovation, is funding the research as part of its Data Exploration program, which focuses on identifying businesses opportunities presented by big data.
The U.S. Federal Election Commission (FEC) launched a new website to improve the usability and accessibility of data about political fundraising and campaign spending. The website, called betaFEC, utilizes the OpenFEC application programming interface that the FEC built in summer 2015 and allows users to search and analyze data about campaign finance. The website also has an election comparison tool that allows users to compare data about how candidates for a particular race are spending money.
IBM has partnered with the Weather Company, which operates the Weather Channel, to apply its Watson cognitive computing platform to the Weather Company’s large network of weather monitoring technology, ranging from airplane-mounted sensors to smartphones that collect weather reference data. IBM will use Watson to try to improve the Weather Channel’s weather forecasting and better predict the chances of natural weather disasters.
The UK government has launched Agrimetrics, a center of excellence devoted to developing strategies to use big data to improve the food and agriculture industry. Agrimetrics, funded by Innovate UK and operated by a group of university and research institutions in the United Kingdom, will work with industry partners throughout the food supply chain to develop and implement data and analytical tools that can improve productivity and efficiency, build resiliency, and improve the sustainability of the food system.
The Obama administration announced a series of measures to improve how the United States monitors, predicts, and protects against space weather—solar activity such as solar flares— which threatens electrical infrastructure on Earth. The U.S. Air Force will work with NOAA to make space environment data publicly available to validate and improve forecasting efforts. The Obama administration launched its Space Weather Data Initiative to publish government data that can aid in disaster response and recovery from a space weather event. The Office of Science and Technology Policy has published its National Space Weather Strategy and National Space Weather Action Plan to guide federal activity in building resilience to space weather.
Mitsubishi Electric has developed a machine learning algorithm that can detect when drivers become distracted and warn them about their dangerous driving. Mitsubishi’s algorithm analyzes data about a driver’s steering, heart rate, and facial appearance to identify when a driver is not paying attention to the road due to a cognitive distraction, such as absent mindedness. Mitsubishi says its algorithm could help reduce driving errors by 66 percent compared to existing distracted-driving detection software that only relies on monitoring a driver’s facial and eye movements to identify if the driver is not looking at the road.
Andrej Karpathy, a researcher at Stanford University’s Computer Vision Lab, has developed an algorithm called deepselfie that utilizes an artificial neural network to identify qualities that make for good selfies and has created a Twitter account that allows users to send selfies for an automated analysis. Karpathy trained the algorithm on 2 million selfies he collected from the Internet to identify popular traits, such as the use of image filters and a slight head tilt. Users can tweet a selfie of themselves to the Twitter account “@deepselfie” and receive a numerical score based on how well their picture utilizes popular characteristics.
Image: The White House.