Big Data and Big ideas?
Asian studies specialists will receive a taste of computational humanities in the upcoming Association of Asian Studies conference.
There are few academic disciplines that appear more incompatible with big data techniques than “area studies.” Defined as a focus on the histories, cultures and institutions of a world region, area studies places great value on the type of detailed knowledge that can only be learned from a period of in-country immersion. By contrast, big data can often seem far removed from the social reality it seeks to describe – bite-sized chunks, cleansed of the messy contradictions of everyday life for processing by computers.
A street scene from Indonesia and the computer hub of the National Center for Supercomputing Applications. To what degree can computational techniques represent the often chaotic complexity of human relations?
The Elite Network Shifts (ENS) project automatically extracts thousands of political elite names from over one million digitised newspaper articles to study patterns of elite behaviour during periods of political change in Indonesia. To mitigate the mis-match between the immensely complex reality of Indonesian political relations and the highly structured computer processes that have been created to contain them, the project constantly employs “ground truthing” measures to test them. From involvement in the many methodological decisions of the extraction and network techniques, to asking elites themselves the extent to which our networks represents their social reality, the country knowledge of our project’s Indonesia experts is integrated into the project results at every opportunity. To share and discuss such challenges, the project team has recently had two panels accepted on “Big Data in Asian Studies“ at the prestigious Association of Asian Studies conference in March 2014. With over 3000 participants and 380 panels, the annual AAS conference will be the perfect testing ground for a discussion of computational methodologies with area studies specialists.
The first panel will bring together researchers from Malaysia, the US and Canada who are engaged in computational extraction and analysis of data from online social media. The second panel focuses on the use of unstructured digitised texts, with contributions from researchers working on collective violence and speeches by political elite, in addition to presentations by Vincent Traag and Jacqueline Hicks on the ENS project.
Broader questions will also be discussed, such as identifying the types of productive debates in Asian studies that are amenable to resolution with the aid of big data. We intend in particular to look at the practice of collaboration among specialists in natural language processing, data mining, and social sciences on e-humanities projects: How are research tasks decomposed into mutually compatible chunks among the disciplines? How is “success” defined in such a collaboration? How can collaborations – for example anthropological “ground-truthing” and big data mining – add value to each other?
We’ll look forward to reporting back to the e-humanities group in the Spring on the reception we receive from the Asian Studies community!