Columbia will offer a new course on how to interpret and evaluate the impact of data next semester in the hopes of facilitating greater understanding of how data is used.
History professor Matt Jones and applied math and physics professor Chris Wiggins announced the course at an event on the role of data science on Monday. The class will begin as a small discussion section under both the history and applied math departments, and the professors plan to eventually expand the course to lecture size.
Both professors said that the idea stemmed from a fear that while governments and large corporations are gathering more data, the public’s understanding of the way in which that data is used is not sufficient.
“When Facebook says your news feed is an algorithm now and everybody says, ‘OK,’ that’s worrisome, because people do not appreciate how much subjectivity engineers put into their algorithm,” Wiggins said.
Wiggins compared the purpose and scope of the class to that of the Core Curriculum, arguing that data is becoming a huge force in modern life.
“What we’re seeing today is a real transition in the ability of data to impact the world,” he said. “We’ve done a great job over the last 100 years with thinking about what every citizen should know about the Greeks, but in the next century I think there’s a need for somebody to think through what every citizen needs to know about data.”
Jones explained that the course will accomplish its goal of providing a general education on data issues by including students from all disciplines in the same class.
“What we’re trying to do is find something that’s in the middle and sets a context that’s needed by both technologists and non-technologists. We’re not going to be teaching basic technological concepts that technologists already understand; scrape that out of the curriculum and you still have a lot that everybody needs to know.”
One goal of the course is to teach students to evaluate claims based on data that has been interpreted by algorithms.
“If somebody says this algorithm exists, therefore you should believe in it, you should be critical of it,” Wiggins said. “Rhetorical literacy is recognizing that if somebody says this algorithm is true, somebody says that to you because they want something.”
Jones said he also hopes that the course will explore more political questions, explaining how expertise in history is useful to understanding why approaches to data collection and interpretation were established in their original forms.
Jones noted the original purpose for the introduction of modern statistical methods as an example of interesting background.
“We’re going to begin with classical statistics and teach all of the technical rigors that go along with that, but never neglect that the key context for that work was eugenics,” Jones said. “The science can be made independent of that context, but it’s important.”
According to Jones, this shows how a historical approach to interpreting the development of science can create new understandings.
“When you do the history of science and technology, when someone introduces something new— whether it’s physical sampling or randomized trials—it’s always resisted, and those moments of resistance are really powerful because they remind you about the way in which you’re making real choices about what to look at and what not to look at,” he said.
Ultimately, Jones cast the advent of the class as in line with Columbia’s liberal arts focus.
“Liberal education is about empowering people to reflect on the conditions of citizenship, and data needs to be a part of that conversation,” Jones said.