A Framework for Facial Image Analytics Using Deep Learning in Social Sciences Research

Abstract

Recent advances in artificial intelligence have provided exciting new tools for scientific research that have been applied in many domains, including medicine, engineering, finance, the physical sciences, and online consumer analytics. Although text analytics is now quite a mature approach, image analytics is still developing as research approach and its use is not widespread in the social sciences, particularly owing to the specialized knowledge required in visualization and machine learning. In this paper, we demonstrate how facial image analytics research can be expedited in the social sciences by introducing a simple research framework for integrating facial image classification data into studies. We explain the steps of the framework, including data collection, processing, collation, and analysis. This is accompanied by an example piece of research examining the influence of hosts’ facial expressions on review scores in Airbnb.