Blomenstock Response

In this article, Blumenstock describes important aspects between data science and its applicability to solve the human poverty problem. Specifically, he highlights the promises, pitfalls, and prospective solutions associated with such topic. As far as promises go, data scientists are beginning to use big data to connect the impoverished with resources in the same way they would advertisements to daily consumers. Examples range from tapping into calling data of mobile phone users to distinguish the wealthy from the poor to using satellite imaging to analyze the affluence of a certain area. With such powerful data, government agencies will now be able to deliver in aid in a more informed and timely manner. Blumenstock proceeds to juxtapose the promises with the various pitfalls. A pitfall for integrating these methods is the unanticipated effects. The incorporation of these methods are merely advantageous to those in power because only a few, profit-seeking companies have the capacity to derive such data, marginalizing the needs of the impoverished. Another pitfall is the lack of validation associated with the data. Companies are quick to incorporate data collection methods without adequate testing, generalizing the data rather than looking at specifics. A final pitfall is the biased algorithm underlying these methods. Mobile phones require electric power and digital platforms require accounts and some degree of literacy. Thus, truly disadvantaged people tend to be under represented in the data. Blumenstock ends by giving solutions to address these problems. One solution is to customize the data methods, tailoring to the contexts for which they are being used for. Another solution is to deepen collaboration. Given that the much of the innovation is currently confined to the private sector, fostering collaboration between data scientists, governments, and the people in the country in question would help make these processes more transparent and ultimately effectual in the long run.

In response to “Good intent is not enough in data science when dealing with the problems which determine people’s experiences.”, I believe it is also very important to assess the mechanism by which this intention is rolled out. Good intention always leads to good methods- those that are contextually thoughtful and can easily be integrated in each environment. In response to “Transparency is the underlying issue to many of these problems, so an increase in this on both ends (data-based issues & human-based issues) could lead to better results.”, I believe it is important for data-based organizations to initiate this transparency because the impoverished already have so much to deal with. This begins when these organizations blatantly state both their intentions and methodologies behind their data-collection to their test subjects. In response to “In lieu of such drastic potential for promoting applications yet demoralizing hinderances, the balancing act can become difficult.”, I believe that the balancing act becomes less difficult when organizations understand the underlying reason for pursuit: to help those in need. When organizations put the needs of other over their own personal gains, they are able to have a true, lasting impact on society.