Producing Statistics for Urban China

Photograph of Charles Chang

Dr. Charles Chang

We are pleased to announce that Dr. Charles Chang will be speaking at the “Deep Dive into Digital and Data Methods for Chinese Studies” series on February 22 and 23. He will introduce a computational method that overcomes the barriers of absent official urban statistics in studying modern China. He will also offer a workshop on geocoding Chinese addresses. A graduate of the University of Wisconsin-Madison, Dr. Chang is currently a Research Associate at the Center on Religion and Chinese Society at Purdue University.

The “Deep Dive” series is co-sponsored by the Lieberthal-Rogel Center for Chinese Studies and the Asia Library, and is co-directed by Mary Gallagher (Professor of Political Science and Director of LRCCS) and Liangyu Fu (Chinese Studies Librarian, Asia Library).  

Lecture: The View from Above: A Computational Method in the Study of Modern China

Time: Thursday, February 22, 2018 3:00 pm - 4:00 pm

Location: Hatcher Gallery (Hatcher Graduate Library Room 100)

Light refreshments will be provided.

In modern authoritarian China, urban statistics may not be available or are available but inaccessible to the public. In this lecture, Dr. Chang introduces a method called “the view from above” that overcomes these barriers such that even in the absence of official urban statistics and even where available city maps have on them only outlines, maps that show detailed land use can still be produced. He accomplishes this using two procedures, integrative scaling and inferential digitization: the one integrates information from one scale to information at another scale, and the other infers land use from massive indirect data, including those from social media and e-commerce websites. He uses Kunming—a medium-sized Chinese city—to demonstrate how the three-step method works. With each step, mapping of the city attains finer detail and scale, ending with the third step, which shows residential communities of different socio-political status—gated communities, work-unit communities, and urban villages. In the larger picture, this method is an example of how digitization and an appropriate methodology can fill information lacunae caused by either lack of resources or deliberately for reasons of security, as in authoritarian states. 

Workshop: Geocoding Chinese Addresses Using Python

Time: Friday, February 23, 2018 12:00 pm - 1:30 pm

Location: ScholarSpace (206 Hatcher Graduate Library)

The room is equipped with laptops, but please feel free to bring your own.

Light lunch will be provided.

The workshop requires REGISTRATION.

           Any scholar in the humanities and social science who studies China faces a wealth of spatial data that carry addresses but lack precise locations, which handicaps our analytical capability. What makes the task all the more daunting is the data’s variety, which ranges from historical events to the news announcement and social media tags of our time. A computational method is needed to automatically convert the textual addresses to numerical latitudes and longitudes (viz., geocoding). A major purpose of this workshop is to present one of these methods. Note that although the workshop is designed particularly for scholars who focus on China, its methods are also applicable to other non-Western countries. Scholars whose research interest lies in, say, Southeast Asia or Africa are welcome. No programming background is required.

In this workshop, we will use a programming language (Python) to geocode a batch of Chinese addresses. It takes three steps. First, we will discuss the concept behind this process and different options of APIs and algorithms. Second, we will use Python to correct the geographical distortions that the Chinese government added. Third, we will use satellite images in Google Earth Pro to validate the results.

Speaker Bio:

Charles Chang earned his Ph.D. from the University of Wisconsin-Madison in 2016. He was elected as the 2016-2017 Postdoctoral Fellow in Chinese Studies at Stanford University and now serves as Research Associate at the Center on Religion and Chinese Society at Purdue University. His research focuses on political communication in contemporary China.

With the rise of the internet, smartphones, and social media, social science is becoming increasingly computational and often involves the collection and analysis of massive data. One type of study Dr. Chang has undertaken is to see how Chinese internet users, who number more than 700 million, respond to an unprecedented political event, for example, the most recent announcements of corruption at the highest level of government. He pioneered a method of graphing such that he could follow citizen response to each official announcement of the political event at a micro-spatial granularity and, time-wise, from day to day and even from moment to moment. Another type of study required him to solve the problem of scarce or inaccessible data in China, an authoritarian state that limits the amount of information it makes publicly available. Starting with maps that offered little useful information, Chang pioneered techniques that enabled him to fill in the gaps and show urban land in considerable detail, including the location, size, and compactness of different communities, which in turn gave him the means to gain knowledge of the social status of the people within each, as well as how they communicate and behave.

At Stanford, he taught a seminar on the methods of “computational social science” and “digital humanities” to graduate students in which he explored how they can be used to study China. His present and future projects include a book that introduces computational social science to a broader social science public, and, more specifically, the use of its methods to the understanding of religion in China, a project that takes him closer to the humanities.