Detecting interdisciplinarity in top-class research using topic modeling

The paper applies topic modeling to the collection of ERC-funded proposals, interim reports and relative publications, with the aim of measuring in a novel way the degree of interdisciplinarity and addressing several open research questions which broadly aim at understanding how environmental conditions can favour the blossoming of interdisciplinarity. Without venturing into potential interpretations and explanations, we present a series of quantitative results linked with the above questions, while deliberately maintaining a descriptive attitude.