This week’s list of data news highlights covers August 24-September 1 and includes articles on a new “Big Data Institute” and a data-driven geological discovery.
1. The Sad State of U.S. Chemical Accident Data
In the wake of the chemical accident that killed 15 and injured more than 300 in the Texas town of West earlier this year, a team of investigative journalists set out to document the state of U.S. chemical accident data. They found that the data is wrong or incomplete nine times out of ten, and simple questions such as “how often do serious industrial chemical accidents occur nationwide” cannot be answered consistently by the available data. The comprehensive article probes several possible administrative and legislative reasons for the poor state of the data.
2. Towards A Data-Driven U.S. Open
The U.S. Open is in full swing, and so is the U.S. Tennis Association’s (USTA’s) partnership with IBM. The partnership, which has been in place since 1990, was renewed for another four years this year. The computing giant has brought its considerable predictive analytics expertise to the event, taking in social media tracking and player statistics to gauge interest and prioritize programming around particular matches. IBM’s approach also includes a sophisticated system for managing data flows within the USTA’s IT infrastructure.
3. The Benefits of Employee Monitoring, Measured
New research from an interdisciplinary team of researchers examines the impact of information technology monitoring on employee performance and conduct. The paper, which was published over the weekend, measured employee monitoring at five chain restaurants, and found that employees who knew they were being monitored both behaved more ethically and drove more sales.
4. Crowdsourcing and Analytics To Help Refugees Access Services
The United Nations Refugee Agency (UNHCR) has hired analytics and innovation management firm Spigit to help improve access to its service among refugees. The company helped UNHCR implement a crowdsourcing initiative to generate ideas about improving the organization’s access and logistics protocols. Using the company’s analytics platform, UNHCR identified trends and recurring complaint issues, and using its “innovation management platform” it developed a roadmap to implementation of these suggestions.
5. NASA Data Reveals Canyon Under Ice Sheet
Scientists have leveraged data from a NASA airborne science mission to map out a previously unknown canyon under a Greenland ice sheet. The canyon, which is longer than the Grand Canyon and equally deep in some spots, was detected through signal processing of thousands of miles of radar data. The discovery was also an exercise in extensive data cleaning; in order to get all the necessary information, the scientists had to pull in datasets collected by researchers from Germany and the UK in addition to the NASA data.
6. Open Data for Voters In Africa
A new open data-driven voter information application is expanding through Africa. GoToVote!, a product of data journalism project Code4Kenya, was used by thousands of voters before Zimbabwe’s recent election to find polling place locations and other information. The app relies on information from the African Open Data site, and is itself open sourced on GitHub.
7. University of Virginia launches Big Data Institute
The University of Virginia announced a new research initiative called the Big Data Institute this week. The institute, which will foster interdepartmental collaboration on organizing and streamlining data collected at the university, is the latest of a number of such programs that have cropped up in recent months, including NYU’s Center for Data Science and the UC Berkeley School of Information’s Masters in Data Science initiative. The university’s administrators hope the center will serve as a model for future research institutes.
8. Predicting When Gamers Get Burnt Out
San Francisco’s Playnomics offers a variety of analytics for free-to-play mobile and social game companies, but predictive churn modeling might be the most unique aspect of their product. This approach, highlighted in a new interview, calculates the likelihood a given player will abandon the game, along with the predicted number of days the player will stay. Then it provides recommendations for engagement with the most at-risk player populations, which game companies can use to prioritize their marketing efforts.
9. Geospatial Intelligence In Africa
Satellite company DigitalGlobe has created an interactive map documenting humanitarian crises, illegal activity, food and water security issues and other major events across Africa. It also gives users a glimpse of the predictive geospatial analytics the firm provides, which are focused on pinpointing areas where certain events are most likely to occur. The target audience includes NGOs tracking humanitarian crises and law enforcement agencies tracking crime.
10. Identity Analytics at the Huffington Post
Following the Huffington Post’s decision to do away with anonymous comments, the online publication’s tech team will now focus its efforts on making sure nobody can game the authentication system. CTO John Pavley expressed the possibility of implementing a third-party “forensic analytics” solution, which could profile commenters and flag outlying behaviors that might indicate attempts to fake identity.