10 Bits: The Data News Hot List
This week’s list of data news highlights covers October 18-24 and includes articles about a New Orleans effort to use big data to reduce the local murder rate and a new Pandora tool artists can use to learn more about their fan base.
Over the past two years, the murder rate in New Orleans has dropped precipitously. Some experts credit the city’s innovation delivery team, the mayor’s data-driven social services team that was inspired by the approach championed in New York City under former Mayor Michael Bloomberg. The team, made up of data analysts, law enforcement, and academic partners, crunched data from multiple departments to identify the best ways for the city to reduce its murders. The team’s insights led to the creation of a gang unit that targets crime hotspots and more funding for illegal firearms seizures.
The United States Agency for International Development (USAID) released a new open data policy last week. The policy established a data library that will serve as a repository for the agency’s data, mandated that the agency and its contractors submit data sets to the library, defined a metadata protocol, and put forth new data access provisions. The agency is using online software collaboration site GitHub to gather public feedback and prioritize new data releases.
Berkeley, CA-based startup Captricity uses image recognition algorithms and crowdsourced manual image processing to convert large amounts of paper documents into data. Tens of thousands of humans, contracted through Amazon’s Mechanical Turk labor service, are paid to double-check the algorithms’ accuracy and provide feedback to improve the algorithms over time. The company, which presently works primarily in the insurance industry, also has a charitable wing in which it provides nonprofits and other mission-driven organizations with data processing services at cost.
Streaming music service Pandora launched AMP this week, an analytics and visualization tool designed to help musicians glean insights from the company’s trove of listener data. AMP, which resembles a traditional business analytics tool, shows artists their most-played songs, where their audience is most highly concentrated, and what demographics are represented among their listeners. The tool is available for free to the more than 125,000 bands whose songs are played on Pandora. The company hopes the tool will help artists have a better understanding of their fans and book concerts accordingly.
Organizations working to fight the Ebola outbreak in West Africa have had to plan their interventions without the aid of detailed maps, making it harder to ensure that every village has been checked for the disease. A team of volunteers is attempting to make the process easier with crowdsourced contributions to OpenStreetMap, the self-described “Wikipedia for Maps” that anyone can edit. The Humanitarian OpenStreetMap Team is looking at satellite images and manually identifying roads, buildings, bodies of water, and other features that can be used for various purposes by emergency responders.
San Jose, CA-based company Xactly makes software that offers companies algorithmically-generated recommendations on how much to pay their salespeople. The company recently released an analysis of its data on women in the workplace, finding that among its customers, women generally outperformed men but made lower commission rates. Xactly analysts then turned the lens to their own company and found that they were incorrectly paying several of its female salespeople. As a result, two female Xactly employees received a raise, while one received a pay cut.
Consumer genetic testing company 23andMe is integrating a large set of family tree records into its genomic ancestry database in hopes of providing a more accurate picture of their customers’ ancestry. The new data, from genealogical website MyHeritage, consists of over 5.5 billion historical records. The companies hope their partnership will create a better product than either one had individually, discovering geographic origins of distant ancestors with the genomic data and using family tree records to make sense of the process by which they are related to modern day individuals. The integration of the two data sets, and an accompanying search service, will be complete by early 2015.
The Vatican Apostolic Library in Rome is digitizing its large collection of ancient religious manuscripts and making them available publicly online. The library, which holds over 80,000 manuscripts along with drawings and notes by famous artists and scientists, will release the data in a format developed by the National Aeronautical and Space Administration (NASA) to store images. The library plans to have digitized around 4,000 total documents by 2018. It hopes to digitize all its holdings eventually.
UK scientists are using drones and connected devices to map areas that are at risk of malaria outbreaks. The drones take photos of their surroundings to identify areas of deforestation, which may indicate areas where humans have recently encroached into mosquito-rich habitats. In addition, the team has fitted macaques in the area with global positioning system (GPS) collars to track their movements, as the disease can also be spread from monkeys to humans.
Palo Alto, CA-based startup Wit.ai wants to bring voice recognition to the Internet of Things. Its natural language processing service, which it offers free to companies who agree to share their user data with the Wit.ai community, allows hardware and software makers to easily add voice control to their products without having to make the large investments required to build ad hoc voice recognition systems. At the moment, around 4,600 developers use Wit.ai to add voice capabilities to mobile apps, robots, smart home devices, and wearables.
Photo: Flickr user Erin Benson