Making corpus data visible: visualising text with research intermediaries

Researchers using corpora can visualise their data and analyses using a growing number of tools. Visualisations are especially valuable in environments where researchers are increasingly expected to work with public-facing partners under the auspices of ‘knowledge exchange’ or ‘impact’, and corpus data are more available thanks to digital methods. But, although the field of corpus linguistics continues to generate its own range of techniques, it largely remains oriented towards finding ways for academics to communicate results directly with other academics rather than with or through non-experts. Also, there is a lack of discussion about how communication, motivations, and values also feature in the process of making corpus data visible. These factors are arguably just as influential for the shape of the final product as technical aspects. This paper aims to open up this process by reporting on two corpus-based projects about press portrayal of migrants led by the Migration Observatory at the University of Oxford. This organisation links university researchers with non-academic users such as policymakers, journalists, and civil society organisations. Comparing the visualisation outputs of these projects, as well as the rationales for key decisions throughout their creation, suggests non-technical lessons for anyone wanting to visualise text: consider the aims and values of partners; develop communication strategies that acknowledge different areas of expertise; and link visualisation choices with wider project objectives