Decolonising Data Science Education

Existing data science educational material represents a narrow view that might unfairly privilege certain demographics while excluding other identities and issues.  

Starting with their own course, Associate Professor Trivik Verma and Assistant Professor Juliana Goncalves aim to decolonise data science education. They created an open interactive textbook that critically examines the way analysis and design impacts or perpetuates issues of inclusion, inequality, participation, power, and positionality. They also developed a formal approach that can be used by others.  

Unequal and sensitive subject 

Understanding the inequalities and vulnerability of communities in urban spaces around the world is central to the research that Goncalves and Verma do. From Brazil and India respectively, they see a lot of inequalities between countries in the Global North and Global South. However, there is very little data available to study the impacts of these inequalities in the Global South. “It’s difficult to digest that we are looking at all of this data without any reflection on what's going on and thinking of data technology as the ultimate equaliser and an objective tool to deliver solutions,” says Verma. “Data is also used to cause injustices in society.” Technology alone can’t solve the world’s problems. He says the solutions are found in understanding how politics, capital, and social realities intersect with technology.  

“I think we have to engage with those debates in a very academic way as well.” Verma and Goncalves run the Centre for Urban Science & Policy together at TU Delft, where they see a lot of resonance on their ideas from students. “We engage with them and ask their opinions and they are often even more critical than we are,” says Goncalves. “They also see these problems and expect us to address them, but we would like to work together with them. I think they must be at the forefront of the change in education.”  

Towards decolonisation  

The first stage of their project involved an ethnographic analysis of the TU Delft community, conducted by Tian Qing Yen, a communications master’s student at the university. “The idea was to study how our community perceived themselves and how that might be related to our own identities and positionality in the world,” says Verma. “We learned quite a bit about who our colleagues are, what they stand for, what they feel, and how they want to engage with the subject of data science and machine learning.”  Goncalves and Verma then started the process of decolonising a course they teach together called Spatial Data Science.  

They wanted to create a process that would be helpful for colleagues who might be thinking about decolonising other courses. The pair developed a formal approach to do that and wrote an article about it, which they hope to publish later this year. In addition, they produced an open interactive textbook which is currently in use for their course. 

Critical impact  

In addition to funding that enabled them to hire people, the project provided validation for the team. “It gave us reassurance that we are doing something that is valuable,” says Goncalves.  “To me, that was a more symbolic value than the funding itself.” And they are getting a lot of positive feedback about their course material from international and Dutch students, both at TU Delft and outside. “It has created a positive space for the student community to feel like they are getting what they wanted from this education,” says Verma. “I’m proud that our students are independent in thinking about these critical issues and that they challenge us and sometimes surprise us with what they can accomplish."