Our paper on the trajectories of scientists in the knowledge space has been accepted in EPJ Data Science! We explore the connections between interdisciplinary mobility and innovation, building on our previous work on the dynamics of scientific fields in arXiv (see this post). Here, we show that tools for studying human mobility can be applied to the mobility within scientific knowledge. Using low-dimensional embedding techniques, we created a knowledge space composed of 1.5 million articles in physics, computer science, and mathematics. The analysis of individual researchers’ publication histories reveals patterns of knowledge mobility similar to physical mobility. Collectively, these trajectories form mobility flows that follow a gravity model, favoring jumps in high-density areas and making long-distance moves less likely. We identify a dichotomy between two types of researchers based on their individual mobility: interdisciplinary explorers, who venture into new fields, and exploiters, who primarily stay within their specialty areas. This work opens up the possibility of using this space to quantify how interventions (such as funding) modify trajectories and the structure of the space.
Here is a thread summarising its findings:
Marc Santolini 