Leveraging Unstructured Data to Address Societal Challenges in the Digital Age

Abstract

This dissertation is a culmination of four years of research focusing on leveraging unstructured data to address societal challenges in the digital age. This work applies and, when necessary, develops data science methods to address pressing societal challenges, including hate speech, the propagation of misinformation, and real estate price appraisals. Additionally, it demonstrates the untapped potential of user-generated online data and imparts effective strategies to harness this resource. The scope of this dissertation covers four papers, each published in renowned scientific conferences and journals, addressing discrete aspects of applying unstructured data to address a pertinent societal issue. Each of the included papers leverages state-of-the-art methods in their respective domains, including dictionary- and machine-learning approaches, successfully addressing the highlighted issues and underlining the untapped potential of user-generated unstructured data available online.