Machine Learning in Social Science

This open access book explores how machine learning can enhance both quantitative and qualitative research in sociology. By developing algorithms tailored to specific data, machine learning enables social scientists to uncover patterns, generate new theories, calibrate indicators, and strengthen causal inference. The book offers an accessible introduction to the principles and applications of supervised and unsupervised learning (Part I), followed by empirical case studies across key areas of sociological research. In the social prediction section (Parts II–IV), it illustrates how supervised learning can 1) impute missing indicators, 2) derive theories directly from data, and 3) improve causal inference through counterfactual construction. In the culture modeling section (Parts V–VI), it shows how unsupervised machine learning can map the structure of large-scale cultural texts—such as online novels and film databases—making complex cultural patterns visible across time and space.

Yunsong Chen is Changjiang Distinguished Professor of sociology at the Department of Sociology, Nanjing University. He earned a D.Phil. in sociology from University of Oxford, Nuffield College. 

Zhuo Chen is Postdoctoral Research Fellow in sociology at the Department of Sociology, Nanjing University. She earned a Ph.D. in sociology from Nanjing University.

Wen Ma is Research Associate at the School of Journalism and Communication, Nanjing University. She earned a Ph.D. in sociology from Nanjing University.

Guodong Ju is Postdoctoral Research Fellow in social attitudes at the China Institute, University of Alberta. He earned a Ph.D. from London School of Economics and Political Science (LSE).