This week’s list of data news highlights covers October 12-18 and includes articles on data-driven dream analysis and a major English effort to consolidate healthcare data.
Developers of the crowdfunded app “Shadow: Community of Dreamers” aim to collect reports of users’ dreams to create a searchable online database for dream analysis. The data, which will be presented anonymously, will consist of typed or spoken dream summaries that users record immediately upon waking. The creators of the app believe that the app could ultimately prove useful to medical science by providing a way for researchers to analyze nightmares and sleep disturbances.
In early 2014, England will begin a comprehensive health data consolidation effort to link previously siloed data from general practitioners and hospitals. The government will then be able to mine the anonymized data to identify and prioritize areas for cost savings and service improvements. Patients can opt out of the program.
Kenya’s open data portal, launched in 2011 to much hype, has proven to be a disappointment, with traffic declining and no updates in eight months. The platform’s stagnation is due to a number of factors, including the Kenyan government’s unwillingness to share information and a new constitution that has forced officials to spend much of their time adjusting to new roles and processes. The portal could still be a success if the country would adopt a freedom of information law and update government data sharing policies.
An online predictive analytics system is being used to help students pick university classes based on their usefulness for the student’s major. The “Degree Compass” program, created by Tennessee-based education technology company Desire2Learn, tailors class rankings to students with the help of data such as past grades and test scores. The program has been implemented in seven universities and community colleges in Tennessee.
A nonprofit working to enroll Camden, New Jersey residents in health insurance under the Affordable Care Act is using publicly available hospital data to map areas of the city with low insurance coverage. Hoping to prioritize neighborhoods by greatest need, the Camden Coalition of Healthcare Providers realized that certain hospital bills could be used as a proxy for uninsured patients. In the future, the Coalition hopes to get hospitals to automate submitting their data, so that such maps can be generated dynamically.
The City of Oakland passed its Open Data Policy this week. The city, a major hub for civic hacking, already has a sizable open data portal, but the policy will be the first time that the goal of releasing public data in open, machine-readable formats has been codified into law. Advocates hope the policy will lead to better prioritization of municipal services and engagement with the city’s 400,000 residents.
Several trucking fleets have begun to experiment with data science to optimize routes and improve their bottom lines. Applications in trucking include simple cross-tabulations that help identify the drivers who present the greatest churn risk to more sophisticated predictive modeling that helps improve safety and driver performance. Many logistics companies collect large amounts of data, but in many cases it has not been put to use beyond traditional business intelligence.
As the demand for greater processing power begins to outpace the rate of processor improvement, computer scientists are starting to think about what data analysis might look like on a fundamentally different architecture: that of the quantum computer. Novel quantum computing algorithms from simple search to machine learning could dominate their traditional counterparts, although the technology will likely take years to mature.
Premise, a San Francisco-based startup, aims to provide near-real-time price information on thousands of consumer products worldwide, using images captured online and with cell phone cameras. In developed countries, much of this information is online and available for web scraping, but in developing countries things are not so easy; the company has 700 contractors who take photos from store to store in 25 countries in order to create the data set. Premise hopes to deliver the information more frequently, completely and cheaply than the U.S. government’s Consumer Price Index.
A group of influential Silicon Valley investors announced this week the formation of Big Data Elite, an early-stage venture fund and incubator for data-science startups. Major backers include venture firm Andreessen Horowitz and angel investor Ron Conway. The group will offer a six-month program for 10 startups beginning in January 2014.