Frictionless Data for Reproducible Research is an open source project within the nonprofit Open Knowledge International that is focused on reducing “frictions” in scientific data workflows. Over the last several years, we have supported a growing community of users and contributors, and have successfully harnessed this network to help build new tools and improve our software, including several Python libraries. This talk is aimed at open source community enthusiasts and will delve into the lessons we have learned on how to support an open source community, including what doesn’t work, and new avenues we are exploring to further sustain our network.