We at Inform believe that data tells a story, across all industries, and every week we round up the most interesting ones right here. This week: data driving crash test dummies; taking on big pharma; the evolution of smiling.
While traditional crash test dummies can provide data on about 20 points on the body, says Technology Review, a new digital simulation can provide much more detail.
Based on five years’ of data collection on thousands of virtual crash simulations, information drawn from a database of injury research, and a digital model with 1.8 million elements on the human form, a research team at Wake Forest University has developed a digital crash test dummy which can test “a variety of body shapes and sizes and different body positions at the moment of impact,” and can “quantify the risk of bone fractures and damage to soft tissue and organs, injuries unaccounted for by crash test dummies.”
Car manufacturers are finding the data invaluable. While using actual crash test dummies comes late in the design process, manufacturers can use digital dummies very early on and make modifications to improve safety, which in turn cuts down on costs as well.
Deutsche Bank is upgrading their systems to leverage data in order to improve customer service and experience. Such an upgrade will “provide a detailed picture of how, when and where customers interact,” and allow the bank to see previously unseen patterns and gain new insights.
The data is often provided by the customers themselves, such as when and how they log in, products and services they use, when and from where they use the products and services. Insights into such data might help Deutsche personalize services according to customers’ specific needs, identify bottlenecks, and solve problems more quickly.
To understand Earth’s complex climate and make predictions such as how greenhouse gas emissions will affect our future, scientists run climate simulations on thousands of linked supercomputers.
Measuring factors such as the amount of sunlight reflecting off sea ice and how the wind affects ocean currents, scientists have come up with a climate model that shows that if greenhouse gas emissions keep increasing, “the world will look different.” For instance, there will be “very little ice left in the Arctic” and New York might be as warm as Miami.
One data scientist is tackling the big issue of skyrocketing health care costs by taking a look at data around brand-name prescription medicines.
Using Medicare drug prescription data from 2013, he studied the number of times a drug was prescribed, the costs, and generic versus brand name costs, and found several patterns.
One was peer influence — that is, doctors are likely to prescribe the same drugs as their colleagues, which are often the more expensive brand names over generics, although generics have been proven to be just as safe and effective as their pricier versions.
Another trend found was patient-driven demand. Pharmaceutical companies are very good at marketing expensive branded versions as new and better while they’re no better than generics.
Until recently, data mining from photographs has proved to be difficult. The data set is immense, starting from the advent of photography 150 years ago, and the information often “too complex or too mundane” to put it into words. However, a machine-vision approach to data mining developed by a research team at UC Berkeley is changing that.
To test their method, the team tackled a database of American high school yearbook photos from 1905 to the present, and found, among other patterns, an “evolution of smiling.”
Right after the invention of photography, most opted for the more easily held neutral pose similar to that used for a painstakingly painted portrait. But as photography became more popular and Kodak advertised the idea of recording “happy memories,” smiling took over, and people began to say “cheese” over “prunes” when posing for a snap.