Details of training program
The DS421 training program will span two years of a student’s graduate study, with most activities occurring in year two. All trainees will need to have full financial support in year two (from a DS421 fellowship, independent fellowships, or other financial sources) to allow the time for program activities as well as fulfilling requirements of their home department. See the training overview for an illustration of the training program schedule.
- Background courses. All DS421 trainees are required to take at least one advanced undergraduate or graduate course in: 1) Statistics (through basic multivariate analysis methods); 2) Microeconomics, Public Policy or equivalent; and 3) Ecology, Environmental Sciences, or equivalent. Courses that fulfill these requirements will be drawn from existing UC Berkeley courses (approved classes will be identified by May 1). Trainees may also submit prior undergraduate or graduate coursework for approval for each requirement.
- DS421 Colloquium (IB 290; 1 unit; fall term). Weekly colloquium with student and faculty presentations, invited speakers, and/or discussion of current topics. Fall 2015: Wednesdays 2-3:30.
- Reproducible and Collaborative Data Science (Stats 159/259; 3 units; spring term). Introduction to principles and tools for reproducible and collaborative data science, including data curation and cleaning, version control, virtual machines, scripted work flow, hypothesis-driven exploratory data analysis, data visualization, and communication. Students will be introduced to git, Python, R, and LaTeX. The class will navigate a series of problem-driven analyses, focused on case studies and independent projects, leading to reproducible products that allow updated analyses as new data become available. Projects by first year trainees will be presented at the Annual Symposium.
- *Interdepartmental Immersion (spring term). In spring of year one and fall of year two, DS421 trainees will be placed in pairs and immersed in each other’s departments, attending social events, research seminars, lab group meetings, etc. Each student will act simultaneously as host in their ‘home’ department and visitor in the ‘away’ department. Pairs will be rearranged each semester (e.g., an ecologist would be paired with a computer science or statistics student one semester and a social science or policy student another semester).
- *Summer Research Experience for Trainees (summer): In the summer following the first year, trainees will be paired with a mentor (a senior DS421 trainee or other senior graduate student) in a complementary discipline to conduct a discrete independent project and assist with the mentor’s research. DS421 faculty, together with the advisors of the two students, will oversee and approve the summer research plan. Trainees, mentors and supervising faculty will meet weekly during the summer for seminars, research updates, and final presentations. Expected commitment is 50% time during the summer, allowing time for preliminary dissertation research and other commitments.
- Interdisciplinary Research Design and Methods (4 units, fall term). Introduction to experimental design, analysis, and inference in natural and social sciences, and application to interdisciplinary research. Close study of exemplary publications integrating natural and social science methodologies will be combined with assignments to conduct similar analyses using faculty-provided or publicly available data sets, putting into practice reproducible and collaborative data science skills. Case studies used to integrate analytic methods and conceptual analyses. Focus on identifying major open questions and research needs in the field.
- *Interdepartmental Immersion (fall term). Continuing from year one spring, as described above.
- Data Visualization and Science Communication ‘bootcamp’ (first 2 weeks of January): Trainees will participate in a 2-week ‘bootcamp’ focused on creative and technical skills in data visualization paired with training in communication to the media, policy makers, and the general public. The intensive format, at a time when classes are not in session, provides flexibility to bring together students, faculty, and working professionals. We will offer this boot camp in collaboration with the Understanding Global Change project in the UC Museum of Paleontology and local public radio/TV station KQED.
- DS421 Team Project (4 units, spring): The culminating course will be an interdisciplinary problem-oriented team project addressing critical challenges in environment, society, and global change. Teams of 3 trainees with expertise in computer or data sciences, environmental science or design, and social science or public policy, will work in collaboration with one or more external partners on a problem of shared interest to produce a two-part product. Part 1 will be a fully reproducible data acquisition and analysis workflow, testing key hypotheses of coupled human/natural impacts and responses to global environmental change. Part 2 will be a proposed solution or adaptation strategy to reduce negative impacts on society in the future, with a clear evidence-based analysis of the efficacy and feasibility of the proposal. A wide range of collaborative arrangements are possible with external partners, such as: production of an interactive data visualization for the web or public exhibits; serving as ‘beta-testers’ for APIs for open access to remote sensing, environmental, economic, or other data sets; collaboration with federal/local agencies in natural resource management or climate adaptation planning, etc. Seed grant funds will be available on a competitive basis to support data acquisition, analysis and project outreach. Collaborations with these external partners will be critical to incorporate constraints and priorities of non-academic institutions in the design of the project, and to expose trainees to a range of professional organizations. The final projects from this course will be presented at the Annual Symposium, allowing students one of many opportunities to put their communications skills into practice.
- Annual Symposium. DS421 will hold a one-day annual symposium in May of each year for all trainees and faculty, as well as the wider campus community. The symposium will include keynote talks from faculty and leading practitioners from our external partner organizations, presentations of team projects, a career panel, a ‘town-hall’ meeting to discuss successes and future directions for the program, and opportunities for informal discussions among faculty and trainees.
- Hackathons and workshops: We will work with the Berkeley Institute for Data Sciences (BIDS), the Geospatial Innovation Facility (GIF), iPython, Software Carpentry, Data Carpentry, the D-lab, and other campus initiatives to promote informal training opportunities in hackathons, workshops and other settings. Students, post-docs, and faculty will be encouraged to organize and lead hackathons, with support from the DS421 program.
- Professional Conferences: Funding will be available for trainees to attend an interdisciplinary professional conference or a conference outside their primary discipline once during the program.
- Speaker series: The DS421 program will host one prominent visitor each semester for up to a week, including one or more public seminars, shared meals, group meetings, and individual meetings with trainees and faculty. Trainees will select and invite visitors, with a balance of academic and non-academic speakers.
*Not required for students in Master’s programs
Last updated: 2/21/15