visualization: collaboration map

data visualization: mapping collaborations

1st place: Duke Research Computing Symposium, Data Visualization Challenge, Jan 2017


selected press


Times Higher Education

- US academics put interdisciplinary research on the map


Duke Today

- Webs of minds and ideas bind Duke's campus

How does proximity influence collaboration? In the 1970's, work by MIT professor Thomas Allen revealed that the frequency of communication between two engineers sharply decreased as the distance between their desks grew. Nowadays, there are more ways than ever for researchers and scholars to communicate and interact. The goal of this visualization was to examine the geographic spread of collaborations represented in the scholars@duke dataset.

view interactive demo

view larger version

download full resolution (~25Mb; right-click on link, save image as...)

A collaboration is defined here as any 2 individuals working together on one or more scholarly publications (journal article, book chapter, etc...) represented in the dataset. Individuals were linked to locations on campus based on departmental affiliation. The Google Maps API was used to determine distances and travel durations between locations. Combined, this allowed for a novel glimpse at how collaborations are distributed across geographically disparate locations throughout campus.

All code for analysis and visualization found at: scholars@duke github repository