The LSE Data Science Institute (DSI) is proud to support efforts to get women into data science and make progress towards gender equality.
On International Women's Day 2022, our data science spotlight celebrates women's achievement and raises awareness against bias by showcasing the success of a leading woman from our community, Dr Milena Tsvetkova.
A sociologist by training, DSI Affiliate Milena is a role model in the field of computational social science. She uses computational methods such as social network analysis, agent-based modelling, and machine learning to study social phenomena.
Men greatly outnumber women in data science. In this interview with the Women Data Leaders project, Milena explains that this is the result of bias and stereotypes, whereby girls and young women are driven out of STEM fields having been led to believe that STEM subjects are too difficult for them.
Milena's advice to young women is to not give in to this discrimination. Milena epitomises this, having become a leader in social science whose success was recognised by Daniela Duca naming Milena in her 39 women doing amazing research in computational social science. Milena has also been a guest on podcasts such as The Know Show and ResearcHers Code, discussing diversity and sharing her research.
Milena outlines her research below.
Dr Milena Tsvetkova
Assistant Professor of Computational Social Science
One of the core tasks of the social sciences is to uncover, understand, and address inequalities and biases in the social world. Many people, including academics, were initially optimistic about the potential of new technologies, and particularly the Internet, to democratise and equalise the social world. We now know, however, that today's digital world is likely even more, rather than less, divided and unequal. Yet, we can exploit the very same new technological capabilities, digital data, and computational methods to study and tackle these problems.
In my research, I use online data and data science methods to investigate how individual decisions and interactions aggregate to produce, reproduce, or intensify social phenomena. The question that I have been particularly occupied recently is: How is socio-economic inequality reproduced in daily decisions and social interactions? Together with collaborators from computer science, physics, and management, I am approaching this question from multiple angles.
First, we conduct online experiments to study how decisions to cooperate with and reciprocate to others could reproduce inequalities in social networks due to arbitrary distinctions by endowments (think privileged upbringing) or group identity (think gender or ethnicity).
In another project, we use data from crowdsourced contests to show that competition and leader boards reinforce skill inequalities because they motivate those who are already performing well but demotivate those who are behind. In a third line of research LSE PhD student Yuanmo He and I investigate whether Twitter users' socioeconomic standing in terms of education, income, and occupation can be inferred from the cultural tastes and consumer preferences they reveal in the accounts they follow. Our goal is to then use these estimates to test whether individuals of lower socioeconomic status engage in negative social interactions, disadvantageous social comparison, and risk-taking behaviours that limit opportunities to improve one's lot.