The Importance of Political Networks

Political Networks Conference Welcome
Whittenberger Auditorium
Indiana Memorial Union
IU Bloomington
Bloomington, Indiana
June 28, 2013

Introduction

Thank you, Bernice [Pescosolido].

It’s a pleasure to welcome you to the sixth annual Political Networks Conference and to the Bloomington campus of Indiana University.

The Importance of Political Networks

As you of course know, the visionary work of a number of influential political and social scientists has led to an understanding of the fact that that people’s social circumstances and their social networks strongly influence their political beliefs and behavior.

Better understanding networks is integral to our understanding of political life and to the complexity of human behavior more generally.

In recent years, there have, of course, been advances in theory and methods in your field. In particular, as a computer scientist with a background in artificial intelligence, high-performance computing, and networking, the prospect of harnessing big data and using it for applications in the physical and social sciences—such as a network approach to examine human dynamics—has been a research interest of mine for over 20 years and I know these kinds of applications are of great interest to many of the participants here.

Many of you likely read the essay in the last issue of Foreign Affairs on the rise of big data, using the example of exit polling on election night.1 The authors noted a number of challenges that many of you have had to confront: that collecting data is costly processing and it is also difficult and time consuming, and so we use sampling to infer something about the total population from a small subset. This approach has clearly served us well, as the recent election clearly demonstrated. But we now also have the possibility of extending our understanding using big data, which some of you at this conference have pioneered. We can further explore voting behavior by drilling down to more specific subsets when we collect all of the data, and “big data” is, of course, the ability to analyze at whatever level of granularity you want, or analyze comprehensively all of the data that is generated in some field in a way that simply was not feasible before. While this neither eliminates nor replaces the need for traditional, rich and deep forms of data collection, it has opened up one more possibility to understand the complexity of human behavior.

Of course, we are just beginning to have a national conversation about the more controversial side of big data: that is, in national security. As Kenneth Cukier and Viktor Mayer-Schoenberger wrote in Foreign Affairs, “in all countries, but particularly in nondemocratic ones, big data exacerbates the existing asymmetry of power between the state and the people. The asymmetry could well become so great that it leads to big-data authoritarianism….”2 I am sure that, as political and social scientists, you also have a fundamental concern with this and other potential vulnerabilities and ethical research problems that could be caused by the way in which big data research is harnessed. We look forward to the insights and conclusions that you will bring to these central questions of privacy and freedom. Political science and the other social sciences cannot only help us better understand the processes that generate big data, they can also provide network methods to analyze these new types of information.

So, we are very glad to welcome you to Bloomington as you discuss political networks in an interdisciplinary world. The important work in which you are engaged carries profound implications for democratic civil society, human health, environmental well being, and many, many more areas.

Network Science at Indiana University

As Bernice noted, I am a computer scientist with a background in high performance computing and networking, and I served for 10 years as IU’s vice president for information technology and CIO. Now, as president of the university, my perspective on IT and networking has continued to evolve. Everything we do in IT and networking must always be seen through the prism of the university’s fundamental missions of education and research.

In that regard, I am proud to note that we have developed at IU a great deal of concentrated expertise in Network Science. We are, in fact, home to a number of people who have made major contributions to the development of the field in statistics, the social sciences, informatics, and now, even in the humanities. IU Bloomington alone is home to well over three dozen faculty who work in network and data science across a wide variety of research disciplines. Of course, we also have an urban research campus in Indianapolis and six additional regional campuses around the state. University-wide, we have more than 60 faculty with interest in network and data science. I think if you were to extend the definition of what one regards as network science, the number of IU faculty who work in this area is even bigger.

We are also home, of course, to the editorial offices of Network Science, the first international journal in the field.

As I suggested earlier, many of the recent advances in network and data science have been fueled by improvements in data storage technology and, even more so, in computing power.

This spring, we dedicated one of the fastest university-owned supercomputers in the nation. Named Big Red II, it was the first petaFLOPS supercomputer to be operated purely as a dedicated university resource, and, as a true supercomputer with very low latency high speed inter-processor communications and large amounts of memory per node, it is ideally configured for faculty research based on the analysis of big data.

I should note, in that context, that about ten years ago, I was part of a debate that was happening nationally about the key directions in high performance, large-scale information technology that faculty across the board would need in the future. I had my own views on the matter, but I decided to establish a committee with some of our Academy members and some of our most distinguished scientists across a variety of disciplines, to create a report that would answer this question. The three components that I asked them to look at were computing power, data storage, and network speed and connectivity.

On the first issue, computing power, the answer came back that after the early tumultuous days of high performance computing, things had really settled down in terms of architectures, and it really was just a matter of dollars. If you put in “x” dollars, you put out “y” teraflops. That was less of an issue than it had been earlier in my career. As to high-speed computer networking, by and large, people were happy with the kind of connectivity that was then provided, especially the inter-university connectivity that was provided through networks like Internet2. The committee members unanimously said that the top priority was data storage. What the faculty said in this report was they wanted to be able to get beyond a situation—and it is almost prehistoric to talk about this—where major data sets all had to be loaded off tape—they wanted to be able to have all of the data online—not just gigabytes, but terabytes and petabytes of data. And there were a couple of instructive comments that were made in this report. One faculty member said that he wanted his data to be around not just for his graduate students, but for his graduate students’ graduate students, and that the long-term preservation and curation of large scale scientific data both in the social sciences and the physical sciences was essential to an institution.

Another faculty member said they wanted to be able to store all the information, noise and all, because today’s “noise” is tomorrow’s information. So using compression techniques that resulted in a loss of information was not acceptable. So, we put in place then a large-scale “data parking lot” which is a hierarchical system, but there is a 5 petabyte data storer behind that, and there is a 20-plus petabyte data tape archival facility, all of it though, online, and all of it can be accessed in real time. So, this massive data capability, I believe, should provide data processing capabilities for data intensive science—which is exactly the kinds of fields in which all of you work—for our faculty for the foreseeable future.

Conclusion

I do hope that your time on our campus and at this conference has already been intellectually stimulating and that it will continue to be so. This opportunity to gather with fellow scholars and engage in discussion about the complex origins and consequences of political networks represents an enormously important aspect of the new frontiers of social science research.

Again, welcome to Indiana University Bloomington.

Thank you.

Source Notes

  1. Kenneth Cukier and Viktor Mayer-Schoenberger, “The Rise of Big Data: How It’s Changing the Way We Think About the World,” Foreign Affairs, May/June 2013, volume 92, number 3, 30.
  2. Ibid., 37.