AI-assisted search for digitized publication archives. Fostering the study of historical figures through the use of Natural Language Processing (NLP) and data visualization techniques

Historians typically navigate digital archives in search of key actors (persons, institutions, etc.) and the impact of their actions on history. Cultural archives heavily rely on keyword-based search methods, often yielding numerous but imprecise results across various sources. Consequently, historians face challenges in comprehending the sequential and contextual facts surrounding key historical figures. The Mini-Muse is a preliminary research project that aims to identify programmatic digital methods and data visualization models to enable researchers and students of history to investigate actors and their actions within their historical context. Mini-Muse relies on Natural Language Processing (NLP) algorithms and data visualization techniques. NLP algorithms extract metadata from unstructured textual content and produce structured data about the facts related to key figures. Additionally, data visualization techniques contribute to visually analysing the timelines of every actor and their relationships.The research project adopts a user-centred design approach to ensure that the AI-assisted search addresses historians' needs. The research team is currently conducting interview sessions with historians and heavy users of cultural digital platforms to gather insights into challenges and desired functionalities. The next steps include the definition of relevant and feasible use cases, the implementation of NLP algorithms and the development of a set of experimental web-based user interfaces.Mini-Muse is made in partnership with E-Periodica, a service by ETH-Library, and it focuses on the “Swiss historical magazine” (Revue suisse d'histoire).