Big Data: Is It the Answer To Every Question?

I view the study of Big Data in a similar way that I see the study of economics, which tries to make mathematical and predictable sense of the human race. As Simon Chandler writes in an article for TechCrunch, “...[Big Data] will do little or nothing to alter the inherent subjectivity of the concepts used to divide this information into objects and relations.” Because computers don’t understand human concepts that can’t be reduced to ones and zeros, using databases and predictive analytics has failed in many projects attempting to answer questions about people and society. Economists are able to make exceptions in their models for the uncertainty of human behavior, but also attempts to use vast amounts of data to make predictions using formulas and numbers – using objective means to make subjective assertions.

The big focus on Big Data has been in marketing and advertising for businesses. Using data about past purchases, things that are purchased together, purchase timing, etc. etc. etc. Anything a business can collect information on. Although this has made huge changes in how products and coupons are advertised to consumers, the models that were created to produce those advertisements were created by humans or with major human input. Accurate models have to take into consideration the society and the culture of the person that they are targeting, and many other humanistic factors that can’t be directly translated into code.

When trying to predict when significant events (such as protests) were likely to happen, there were a number of huge database systems that came together to test their hypotheses. To test these systems, like the Lockheed Martin’s International Crisis Early Warning System (ICEWS) and Georgetown University’s Global Data on Events Language and Tone (GDELT), they measures how accurately they were able to predict protests in Latin America. They had mined news articles and social media to find trends, but found that the project ultimately failed with a little over 10% overlap. They largely attribute this to how “Automated systems can misclassify words.” Again bringing in the concept of subjectivity and trying to have a computer understand what is meant in tone, local dialects, and exaggeration.

While I think that Big Data is an incredibly exciting subject and will prove to be much more useful than the pessimists writing about it believe, I do agree that data alone will never be able to translate the human condition into binary. For Big Data to be useful and meaningful to areas that include any kind of human input, there has to be some kind of subjective influence. Until artificial intelligence is advanced enough to do away with human coders or developers, Big Data will be like typewriter keys with all of the paper and ink, and no author. Maggie Wilcoxon.