Replication Data for Chapter 4: The Footprint of Online Gender Narratives: A Cross-Lingual Deep Learning Approach

New techniques have been introduced over the last decade to extract meaning from unstructured text and big data in Economics. These range from statistical natural language processing methods to deep learning approaches (Abramitzky et al., 2025; Dell, 2025; Gentzkow et al., 2019).

Despite growing interest in deep learning methods among Economists, there are several caveats in their application to developing country contexts, such as a model’s capacity to address cultural and linguistic diversity and its utility for sensitive topics such as gender and discrimination. This paper demonstrates how novel methodologies can be applied to multilingual and sensitive contexts, employing insights from unstructured text data.