Building a data science curriculum and community for ocean scientists, engineers, and students using The Carpentries model

Abstract:

Computational reproducibility in scientific research requires learning and adopting new skills, often easier when learning with peers. To enhance data science and computational skills for our students and staff, Woods Hole Oceanographic Institution has joined The Carpentries, a worldwide community teaching foundational data and coding skills. Until this year, our workshops were taught mainly by external Carpentries instructors, who we tried to select to match with our audience and learning expectations. However this proved to be quite challenging, as our audience ranges from undergraduate students to technical specialists, and our learning expectations are influenced by our desire to enhance skills relevant to the ocean sciences and engineering. This year, six of our staff and students trained to become Carpentries instructors. Never would we have guessed how important this training would be, as our in-house instructors gave us the flexibility to quickly respond to rapidly changing workplace circumstances due to the coronavirus pandemic and offer Carpentries workshops online. Our 3-hr workshop sessions included a workshop for regular expressions, a workshop for tabular data with Python, and two workshop series over 5 days with either Python or R plus shell scripting and git version control. Post- survey results indicated an increase in learner agreement with the statement “I can write a small program, script, or macro to address a problem in my own work.” Based on our experience as instructors and helpers and our assessment of pre- and post- survey results, we infer that workshops taught by instructors from within our own community confer benefits such as domain expertise that may enhance the ability to "teach research with programming" while simultaneously fostering community building. Tailoring The Carpentries materials to our domain improves learning outcomes, but modifying the curriculum for ocean sciences takes time. Teaching online presents new challenges, one of which is building a community of practice without face-to-face interaction. We recommend adopting The Carpentries model to build “capacity in essential data and computational skills for conducting efficient, open, and reproducible research” to other institutions with graduate and undergraduate students in the Earth and life sciences.