Data science teams working on a shared analysis face coordination problems such as dividing up the work to be done, monitoring performance and integrating the pieces. Research on distributed software development teams has raised the potential of stigmergic coordination, that is, coordination through a shared work product in place of explicit communication. The MIDST system was developed to support stigmergic coordination by making individual contributions to a shared work product visible, legible and combinable. In this paper, we present initial studies of a total of 40 student teams (24 using MIDST) that shows that teams that used MIDST did get the planned benefits for their work, did seem to coordinate at least in part stigmergically and performed better on an assigned project.
https://doi.org/10.1145/3449110
The 24th ACM Conference on Computer-Supported Cooperative Work and Social Computing