Text as Data: A New Framework for Machine Learning and the Social Sciences

Text as Data: A New Framework for Machine Learning and the Social Sciences This book provides a comprehensive guide to using computational text analysis to gain insights into the social world. With an abundance of textual data available, researchers can uncover fundamental questions in the social sciences, humanities, and industry. The authors offer a sequential, iterative, and inductive approach to research design, covering core tasks such as text-representation, discovery, measurement, prediction, and causal inference. By combining new sources of data, machine learning tools, and social science research design, researchers can develop and evaluate new insights. The book presents real-world applications, example methods, and a distinct style of task-focused research, bridging divides between computer science and social science, the qualitative and quantitative, and industry and academia. This resource is ideal for anyone looking to analyze large collections of text in an era where data is abundant but challenges remain. The authors' approach provides a practical framework for navigating the complexities of textual data, making it an essential tool for researchers and analysts working across various fields.