|By Haim Koshchitzky||
|July 28, 2014 07:45 AM EDT||
Sharing data is a cornerstone of the scientific method because it makes it possible to replicate work. That foundation is mostly absent from data science, which makes obtaining and reusing knowledge more difficult than it should be.
Job postings for data scientists increased 15,000 percent between 2011 and 2012, and Gartner predicted that 63% of organizations would invest in Big Data this year. The communications, consumer, education, financial, healthcare, government, manufacturing, and retail sectors are all adopting business practices that are using data science to inform their activities and improve operations.
There are a number of companies creating solutions to visualize and uncover insights from large volumes of data with robust platforms in operation worldwide. Vast volumes of data from applications logs to the network and business activities are well served by today's analytics technologies - computation isn't the issue. The ability to model data into experiments that act on data with data sources and conclusions is what's missing, and it's an emerging problem for businesses.
Gartner has observed that those same organizations are now "struggling" with deriving value from and managing Big Data (depending on organizational maturity). That could be due to what famed microbiologist Ludwik Fleck deemed an "empty mind" as he explored the sociology of science during the 1930s. What is that exactly? Fleck postulated that a mind must be filled with initial knowledge before it can perceive or think. This logic applies to organizations too.
Fleck's theory was that participating in a "thought collective" of institutional knowledge would fill minds. His works concluded that cognition is a collaborative activity because a body of knowledge is acquired from a group. It could be argued that making it possible to reuse data experiments would have the same effect. Organizations that can't find value in data have an empty mind.
Big Data should follow the lead of the scientific method (which was influenced by Fleck's ideas) to put greater emphasis on sharing and reusing data. Why is that important for businesses? Scientific data is easy to share among different organizations. Having the ability to do the same with data science could solve what's emerging as a major pain point. Employees change roles and organizations, but what happens to the knowledge, experiments, and patterns?
Whether the academic model would also function in the enterprise is a fascinating question for data scientists, operations professionals and industries. The next great "open source" horizon could be the exchange of knowledge.
It would be interesting to see companies take on the challenge of building systems that organize and share experiments more liberally to put an end to the empty brain problem. After all, data science is still science. Why should it be treated differently?
Jun. 30, 2015 05:00 PM EDT Reads: 2,081
Jun. 30, 2015 04:39 PM EDT Reads: 562
Jun. 30, 2015 04:21 PM EDT Reads: 603
Jun. 30, 2015 03:59 PM EDT Reads: 556
Jun. 30, 2015 02:15 PM EDT Reads: 2,149
Jun. 30, 2015 01:56 PM EDT Reads: 678
Jun. 30, 2015 01:45 PM EDT Reads: 1,995
Jun. 30, 2015 01:45 PM EDT Reads: 1,861
Jun. 30, 2015 01:30 PM EDT Reads: 2,110
Jun. 30, 2015 01:15 PM EDT Reads: 2,052
Jun. 30, 2015 12:30 PM EDT Reads: 1,845
Jun. 30, 2015 10:30 AM EDT Reads: 1,896
Jun. 30, 2015 10:20 AM EDT Reads: 557
Jun. 30, 2015 10:15 AM EDT Reads: 855
Jun. 30, 2015 09:45 AM EDT Reads: 896