While the research discussed here focuses on primary sources related to nineteenth-century Britain, remarks from each presentation will reflect on the ways that working with big historical data has implications across different periods and areas. For example, panelists will review interdisciplinary collaboration, data access, collection and digitization bias, methodological transparency, working with inferred data (e.g. the results of Machine Learning pipelines), or finding suitable audiences for computational historical research. Each presentation focuses on the research related to a particular type of primary source: newspapers, maps, books, and census microdata.

LwM’s ambitions were historical, technical, and cultural: the challenge to use open data to make new claims about the experience of living through industrialization required novel configurations of digital infrastructure and collaborative behaviors. Doing reproducible research in the open isn’t just hard for scientists, it’s also not the modus operandus of historians. LwM therefore stretched - in productive ways - the comfort zones of multiple disciplinary practices when they questioned how future historians would have access to the data and methods being created.

The new research we are able to present in this panel should therefore be understood in the context of the everyday work between historians, linguists, data scientists, research software engineers, geographers, and librarians who were aiming for the same goal: to make our research meaningful to as many people as possible, through its open-access, reproducible insights. This panel aims to explore the relationship between the availability of primary sources, the nature of historical argumentation, research and project design, and the ideals of open research communities, thus far more common in the sciences.

We invite discussion with an audience of colleagues who are working, or wish to work, on computational history; curious about the impact that the open science movement is having on the historical discipline; interested in collections as data; and would like to discuss the implications of using data generated using Machine Learning methods on the significance and trustworthiness of historical claims.