America’s statisticians believe that in the big data age, the most skillful exponents of data – data scientists, domain experts, and mathematicians – should be the most utilized.
In a recent whitepaper, the American Statistical Association (ASA) calls for the inclusion of statisticians in big data research teams throughout the intellectual, science, and business environments.
It is a quietly confident appeal that does not ignore the stark contrasts between academia and commerce. The mission of study can be pursued with very different tempos than that of business. The scholar advances with patience, spending time making every research effort stronger. The marketer, on the other hand, can’t move fast enough for information about the audience.
With these values framing the operational differences between business and study, the whitepaper, authored by a dozen ASA members and data experts, poses a powerful assertion: statistics is the most mature of the data sciences, but is not being used to its full potential in big data research.
Central to the ASA’s view is that a multidisciplinary approach to data usage will bring skeptical insight to bear on any concept, be it commercial or intellectual.
As the report argues, “Insight is required to distinguish meaningful signals from noise. The ability to explore data with skepticism is required to determine when systematic error is masquerading as a pattern of interest. The keys to skeptical insight are rigorous data exploration, statistical inference, and the understanding of variability and uncertainty.”
“I work at a business school and I was trained as a computer scientist. I work with industry all the time,” said Cynthia Rudin, associate professor at the Massachusetts Institute of Technology. “I strongly believe in the transformative potential of interdisciplinary research.
“I think there are a lot of possible education tracks that would allow someone to contribute meaningfully to big data,” she added. “Now there are new specialized tracks being formed in data science, but I think people with degrees in applied statistics, computer science, applied mathematics, engineering, and several other fields all could have valuable skills for working with data.”
A view at various disciplines shows an evolution in the science culture’s attitude toward data.
As the ASA notes, the expansion of genetic research has led biologists to include data usage in the realms of gene, cancer, and related life-science studies. Public safety experts now use predictive analytics to prepare for emergencies or track criminal behavior. Graphs, surveys, polls, and empirical tools are fundamental parts of research for sociologists and political scientists.
When working in business, one of the key skills for a data professional might be adaptability.
“In terms of whether a Ph.D. is needed, it really depends on the project,” said Rudin. “(Sometimes) all you need is logistic regression and a little bit of processing, so no Ph.D. required. I do think a Ph.D. teaches people how to think critically and to have a broad perspective, which can be extremely important when handling a challenging task with lots of difficult-to-work-with data sources, a not-clearly-defined goal, a not-clearly-defined evaluation procedure, and so on.”
The cultural assimilation of data scientists might depend on public relations. “We need to show (businesses) lots of success stories,” said Rudin. “As our community starts to focus more on impact to society, the managers should hopefully get used to seeing the power of what we can do. At that point, we won’t have to do any convincing. Many companies are already convinced. They are hiring data scientists, statisticians, computer scientists, and so on like crazy!”
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise.