Kyle R. Brady: Profile | Blog | Reads

Sunday, April 27, 2014

In Brief: A Look at Big Data and Academia

Cukier and Mayer-Schoenberger, in the May/June 2013 issue of Foreign Affairs, present an argument for the use of so-called big data -- “the idea is that we can learn from a large body of information things that we could not comprehend when we used only smaller amounts” (2013) -- in a world greatly changed by technology, both within academia and the wider world. Beginning with an observation that the digital era has produced seemingly boundless progress in nearly all fields, the authors note that the Internet has allowed greater, more visible, and less faught access to various forms of information than ever before, including previously inaccessible or unidentified data formats. With the ever-greater quantification of information, as well as information made far more explicit and detailed, there exists the potential to assess large datasets on an unprecedented level, searching for previously unidentified associations or to base conclusions more directly upon a wide-ranging sample, rather than a selective one.

This arose to prominence, say Cukier and Mayer-Schoenberger, in the digital era, but it was not an outgrowth of the world’s newfound connectivity, as, for example, IBM “delved into statistical machine translation” (2013) through a preliminary exploration of big data in an early 1990s project comparing official Canadian documents, in both English and French, in an algorithmic fashion. Although somewhat successful, such an attempt was not overwhelmingly so, according to the authors, until Google reimagined this project in their own image -- with their unimaginably large and expansive datasets -- and produced Google Translate, with a much higher degree of accuracy. The Internet, although not a parent of big data, is assuredly the true operationalization of big data.

Through a number of short case study examples, from UPS’ maintenance sensors to human posture monitoring to New York City’s big data experiments, Cukier and Mayer-Schoenberger make a very vivid and positive case for the use of big data to better explore all aspects of human life and increasingly use online “data in the wild” (2013) to expand research beyond the artificial strictures of sampling that intentionally avoids approaching the ideal “n = all” (2013). Reducing a reliance upon random and/or representative sampling, they argue, increases efficacy through sheer quantity and minute, previously implausible, quantifications of various inputs. Although they acknowledge an automatic increase in “some messiness” (2013) of datasets constructed as such, “the insights that a massive body of data provides” are worth the marginally-greater risk, mitigated as it is by algorithms and the auto-correction of error that comes with a substantially increased sample size. A true risk in the era of big data, however, is acknowledged by the authors: “[it] could become Big Brother … exacerbat[ing] the existing asymmetry of power between the state and the people” (2013). This very real possibility would allow the quantification of an individual's life, or a group’s dynamics, to be easily created, profiled, and used for purposes that run counter to the principles of freedom.

The argument that Cukier and Mayer-Schoenberger makes in this essay, although more implicit than explicit, is a simple one: the scientific method, particularly in its more traditional form, is no longer always required. With the arrival of big data, the scientific method is considerably modified, truncated into five all-digital steps: create hypothesis, acquire dataset, process algorithmically, review results, and, if necessary, modify algorithms. No longer must datasets be built through careful representative or randomized sample selection; even the creation of a hypothesis has become optional, as any sufficiently large dataset will produce quantifiable results on its own, following both sorting and analysis, without any prior expectations or goals necessary. These two aspects are greatly innovative, especially since they remove two potential sources of bias -- sample selection and seeking an expected result -- and help to streamline the research process.

As the authors state toward the end of the essay, “[b]ig data is a resource and a tool ... meant to inform, rather than explain; it points toward understanding, but it can still lead to misunderstanding” (2013). The issue of false causation is still extremely germane to any discussion of correlations found within datasets, as can be the correlations themselves, if the dataset is not sufficiently large or the ways in which it is processed are improperly designed. The sources of error, therefore, in any big data approach to research exist in collection and statistics -- computer programming, perhaps -- rather than more human and inferential errors, placing a greater emphasis on technical and mathematical proficiencies than in previous eras. In terms of quantifiable data -- such as actions, events, characteristics, and data naturally numerical -- this is a perfectly rational and reasonable approach.

The question, however, is whether such an approach is altogether appropriate and the answer, without qualification, is in the affirmative. Today’s world is full of keystrokes, data packet transmissions, and geospatial data, as well as freely and openly volunteered personal information. The latter includes both structured data -- name, address, phone number, email address, quiz responses, various forms of favorite content, and considerably more -- and unstructured data -- blog posts, social network updates and inter-personal messages, photographs, recorded audio and video, and many other free-form methods of both communication and expression -- that can all be mined for value, to build effective models -- both descriptive and predictive -- of an individual. Having such information, it would be an inherent oversight of any research project to fail to include a big data approach, even if only for the better understanding of the individuals involved. Moreover, big data can be used to produce research in a much more real-time environment than a traditional approach, as datasets on various research questions can be quickly and rather easily constructed using public, voluntary, individualized, and highly descriptive digital information. In fact, all of this is already occurring, both within the systems the data is created (such as Facebook and Twitter), interested organizations (such as corporate research groups and political groups), concerned governments, and certain portions of academia.

It would be a considerable loss, and the ultimate expression of an anti-knowledge bias, for the academic world to not fully embrace and implement the principles of big data, wherever useful and appropriate. The scientific method is not invalidated by big data; rather, it is augmented for the betterment of science, itself, and is merely the latest evolution of process and procedure. Quantitative data analysis, particularly within the social sciences, rose to prominence some time ago, but its true golden era may still be on the horizon, as its desire to produce scientific results may finally merge with the required technical tools and datasets. An essential element of such progress, however, is understanding when a big data approach -- quantitative data analysis -- does not, yet, apply.

References

Cukier, K. N., & Mayer-Schoenberger, V. (2013, May 1). The rise of big data: how it's changing the way we think about the world. Foreign Affairs.

Saturday, April 5, 2014

In Brief: Presidential Power, Foreign Policy, and Foreign Policy Actors

The foreign policy roles -- roles, too, in the international theatre -- of the various agencies, departments, and advisors are all known to vary in their influence and strength by each Administration. Following the collapse of the Nixon Administration, decades of decreased Presidential control were seen, wherein the foreign policy of the United States was far more dependent upon the various actors contributing to, or carrying out, foreign policy decisions, including Congress. This seemed only the natural result of a President who, privately, may have exerted too much control over the bureaucracy and, very publicly, was shamed for poor decisions that led to regrettable, illegal activity.

However, as of September 11, 2001, this seems to have, once more, changed. President George W. Bush may have originally had a diminished interest in foreign policy or articulable foreign policy goals of any great consequence, but the seminal terror attacks on American soil visibly and demonstrably changed him, his Administration, and his goals. Much can be said about the years that followed -- wherein President Bush pursued two military conflicts in two different states and many other, more quiet, engagements -- but is nearly without question that the agencies, departments, and advisors were made to bend to the will of the President and the Vice President, rather than the alternative. Moreover, interdepartmental, internecine conflict ceased, for some time, as did Congressional opposition, and Presidential authority was once more recentralized in the Executive, although not without ample legal memos, interpretations, and rulings.

President Obama has, to a great extent, continued in this vein of centralized political leadership, albeit with a highly confrontational Congress. Regardless, the agencies and departments seem highly dedicated to the President’s articulated goals and underlying ideals; it is, instead, the Congress that seems intent on opposing the Administration, both publicly and privately. This has led to no small amount of foreign policy confusion -- such as the longterm pro-Iranian negotiation Administration and the recent anti-Iranian negotiation legislation -- but, ultimately, United States foreign policy has managed to remain rather consistent and coherent in its actual execution. Arguments have been made, and may be more clearly demonstrated through the lens of history, that President Obama has established himself as somewhat of a foreign policy President and successfully synchronized, if not centralized, all the various constituent actors.

It is, indeed, the role of the President and the Administration to faithfully execute foreign policy actions that seek the best interests of the United States, rather than any of the agencies or departments, or even Congress. The agencies and departments exist to advise, using all acquired expertise, the President while, simultaneously, carrying out that which the Executive requests: there may be differences of opinion, or even staunch disagreement, but it is, unquestionably, unacceptable to act in direct contravention or circumvention of orders. This is not to say that the President should be allowed full freedom and latitude to undertake an uninformed course of foreign policy, but neither should those agencies and departments of foreign policy find themselves creating foreign policy. Coherence being essential, there should be substantially more longterm foreign policy planning and execution than currently exists, as American interests are unerringly harmed whenever foreign policy goals change wildly with Administrations and bureaucracy leadership; however, this should not be at the expense of a President’s freedom of action or authority. Some form of overall Foreign Policy Doctrine would be appropriate, perhaps, wherein all relevant actors -- excepting the Congress, the form and function of which is wildly mismatched with foreign policy needs -- agree to general goals and action plans, with certain details necessarily provided by each successive Administration.

Tuesday, April 1, 2014

In Brief: Looking at "The World America Made"

It would be somewhat of an exaggeration to suggest that the world-at-large was somehow “made” by the United States, as Robert Kagan does, but it is, beyond any doubt, a world that was heavily influenced by American actions and worldviews. This is particularly true following the collapse and dissolution of the Soviet Union, which left the United States as the sole Great Power -- a universal hegemon, perhaps -- and better able to project its various forms of power. There are, of course, those who question the strength and influence of the United States in the modern era, following post-September 11, 2001 wars in Afghanistan and Iran, the increasing destabilization of various regions, and an apparent resurgence of both China and Russia, with the latter as a possible kindling of Cold War II. Those who undertake such a line of questioning, however, mistake hegemonic power for imperial infallibility.

The world is a very dangerous, dynamic, and developing theatre of threats, but this is not new. All of history’s empires faced threats, within and without, and they were nowhere near omnipotent or without failures; each empire’s collapse demonstrated these characteristics, true, but such events were not exclusive to the process of demise. American infallibility and the multitude of threats it faces should not, therefore, be seen as a negation of its power or presence -- it is, rather, simply an expression of reality.

Having established that infallibility is not a prerequisite of exerting hegemonic influence, it can be observed that the United States is in exactly such a position. Regardless of the threats, however, American power and influence continues unabated and relatively unchallenged, as there exists no strong indication that China, Russia, or any other state is able to provide a match, on any level. The United States still finds itself in a predominant position, in all regards, and many of the challenges mustered by others are either largely symbolic or lacking sufficient impact. That the American political, military, or socio-economic machines may not always muster themselves for a full-throated response is no indication of anything other than a lack of serious or compelling interest in the issue at hand.

The projection of American power has, without a doubt, waxed and waned over the centuries, but it appears no measurably weaker than during the last few decades. The focus of response efforts may have shifted, both in terms of paradigm and result, but it would seem that a realist approach has rightfully returned to the international behaviors of the United States: selfishness and security concerns. For the decades to come, these goals will still be vigorously pursued, in ways determined the most appropriate, but many situations will continue to be ignored or given half-hearted efforts, when they do not meet a sufficient level of self-interest. Some states may incorrectly perceive this as weakness and rhetorically seek its exploitation, but the hegemonic nature of American power will continue to project, unabated, into international arena, always providing an overwhelming response as soon as self-interest dictates.