This week’s list of data news highlights covers November 16-21 and includes articles about using data science to fight crime and detecting heart disease with predictive analytics.
Two groups of scientists from Google and Standford University have created artificial intelligence software capable of recognizing and describing the content of photographs with an unprecedented level of accuracy, sometimes comparable to human levels of understanding. Unlike previous computer vision technology, this software can teach itself to identify entire scenes instead of just individual objects. Researchers have high hopes for the future of this technology, hoping it can one day be used to help the blind and robots navigate natural environments.
Thanks to a new $320,000 grant from the Bureau of Justice Assistance, the Philadelphia Police Department will being hypothesis testing as a means of developing crime fighting predictive analytics. The Philadelphia Police Department had previously been training 30 officers in data science to develop better crime fighting techniques. With hypothesis testing, the department hopes to be able to understand why and where specific crimes occur so they can develop strategies to prevent them from happening. Department leaders hopes data science can become part of the DNA of policing and integrated into everyday practices.
Researchers at the University of Nebraska have developed an interactive mapping tool to increase crop yields to tackle the problem of feeding increasing populations. Making existing more farmland more productive is crucial to producing enough food to feed rising populations without converting millions more acres of land per year to farmland, according to the developers of the tool. The mapping utility, called Atlas, integrates site specific data on soil, climate, and cropping systems in nearly 20 countries around the globe, and researchers are working to secure data for 30 more countries.
Cities hit with snowstorms this past week, such as Boston, Minneapolis, and Buffalo, are using machine-to-machine communication tools to expedite snow removal and help citizens. In some of these municipalities, when a city plow clears an area, a GPS sensor on the plow will automatically close the request call on the city’s computer system, which then emails the resident who made the request that the street is clear. Buffalo’s director of the Division of Citizen Services hopes that the Internet of Things can be expanded to other public sector services to improve transparency and guide decision making on tight budgets.
The Centers for Medicare and Medicaid Services (CMS) has named Niall Brennan as its first chief data officer, responsible for improving the data collection and dissemination capacity of the agency to improve transparency. CMS expects the efforts to improve data transparency will improve outcomes, reduce costs, and better inform consumer decision making. Brennan will head the the newly created Office of Enterprise Data and Analytics.
The Harvard School of Public Health has created a new online tool to help patients discover their risk of developing cardiovascular disease or of having a stroke. The tool uses a predictive analytics model incorporating lifestyle factors that may eventually produce heart disease in otherwise healthy patients. Harvard researchers hope their tool can extend the ability of healthcare provider to deliver preventative, long-term care instead of just treating patients with more immediate concerns.
Following a series of potential police misconduct towards minorities in Durham, North Carolina, a coalition of social activists turned to data instead of protests. The group carried out an analysis of 11 years of state police data to reveal that Durham police disproportionately targeted black male motorists during traffic stops, despite no greater likelihood of finding drugs or other illicit substances. Despite previously not taking action, the city changed course after this analysis was presented to require police to obtain written consent to search vehicles in cases without probably cause. Collection of police activity data is a growing trend around the United States, which activists hope will continue to inspire policy change.
While cell phones have long been used to determine location, which has been very useful to emergency services, advances in cell phone technologies have made it possible to determine a location’s altitude. Barometric-pressure sensors are a standard feature of the newest line of smartphones, with more than 100 million already owned by consumers, and are capable of determine altitude to within a few feet. More accurate location data has enormous potential for public services like emergency response crews trying to reach people in tall buildings.
In San Francisco, teams of data scientists competed in a 24 hour competition to create a social good-themed app in a hackathon hosted by a local nonprofit. Over 100 data scientists, engineers, and designers competed to build the best app that could benefit the public. The winning team developed a web application that detects prostitution rings by monitoring where adult ads are posted online, using data from a nonprofit devoted to fighting child sex trafficking. As hackathons grow in popularity, organizers felt it important to host one specifically focused on data science to develop better products to benefit the public.
Citing the need to address the notoriously poor air quality of Chinese cities, a team of researchers at the University of California, Berkeley have developed a new kind of air pollution sensor. The gadget, named Clarity, fits on a keychain and constantly tracks personal exposure to air pollution, then relays that data to a smartphone. The researchers hope that with widespread adoption of the device, pollution can be mapped on a wide scale in real time to strengthen efforts to improve air quality.