Maxwell Fonss
January 28th, 2020
While the promise of big data may offer substantial improvements to the average quality of life in both first world and developing economies, Joshua Blumenstock notes that there are still many obstacles that prevent this new field of science from being used to it’s maximum potential.
Blumenstock brings up several promising methods in which big data is currently being used to help those in need, including the mining of data from cell phones and satellite photography. While convential methods of determining reliability for loans such as credit scores are unavailible for many in developing countries, big data has allowed algorithims to bridge the informational gap between lenders and borrowers by using algorithms to use unconventional information such as social media content, giving greater financial freedom to millions of people. Similarly, Blumenstock also mentions satallite images being used to by humanitarian agencies to better geo-locate aid to those in need. Using machine learning and image recognition software, organizations can sort through very large volumes of data quickly to determine areas that may be lacking proper food supplies.
However, like all emerging industries, big data faces many unique problems that may not only hamper it’s effectiveness, but actually harm the people it is being used to benefit. Big data is still in it’s infancy, and many data scientists are still learning which independent variables they should measure to draw accurate conclusions. This becomes even more difficult when acounting for the fact that people’s behavior may change as they become aware that certain actions are being measured, or that bad actors may try to manipulate data for their own personal gain. Even if we can perfect our analysis of large datasets to accurately determine outcomes, private interests or totalitarian regimes could still use big data to exploit rather than help those in need.
Despite the obstacles facing big data, Blumenstock makes it clear that big data can be a force for good in the modern world. He offers several solutions for the previously mentioned problems facing big data, such as a more diverse field of data scientists that better understand the problems they are addressing, working with other fields of academics to avoid externalities, and putting more resources into the hands of researchers focused on holding machine learning algorithims more accountable.
Personally, I think that with the proper regulation, private interests will be able to make sure that big data provides the maximum amount of utility to the global community. Most of the solutions Blumenstock suggests will require large amounts of funding which I believe can only be provided by investors.