Capstone Partner FAQ
The Duke Master in Interdisciplinary Data Science (MIDS) Capstones must address an important problem, and allow MIDS students to contribute meaningfully to the solution. All Capstones must have at least one stakeholder (such as a company, a government agency, an internal Duke organization or a nonprofit) who agrees to be available throughout the year to evaluate the project, and at least one Duke researcher with expertise relevant to the problem being solved. Project teams can be as large as necessary to address the problem and can include multiple faculty members, postdocs, students at all levels (graduate and undergraduate), and other staff. Note that even though students work collaboratively in Capstones, each MIDS student must achieve a specific outcome of interest for the stakeholder, and give a final presentation with an accompanying white paper about the implications of that outcome. To ensure MIDS students complete their projects successfully, they attend workshops and complete assignments throughout the second year that provide guidance, practice, and feedback about students’ teamwork, project management, communication plan, and overall progress in relation to the project. They also have a mid-term review and final presentation that is evaluated by a panel of MIDS core faculty and outside partners on multiple dimensions including students’ ability to communicate effectively to a diverse audience, computational strategy, and creativity.
In order to be eligible to host a Capstone team, organizations must be doing work that creates a question or problem that can be addressed via the tools of data science, provide students access to the data they need to address that question or problem and be willing to devote the time (1-3 hours per week) to mentor a team of two to four students. Governmental, non-governmental, non-profit, for-profit, and university-affiliated organizations are all appropriate.
Prospective Capstone partners submit an application describing their project and data, and we work with them to assess its fit with the Capstone requirements and craft it to ensure that it connects with MIDS students’ interests and goals. Once a proposal has been fully develop and accepted, MIDS students and the Capstone partners participate in a matching process; projects do not move forward unless they appeal to one or more students, but with partner approval may be kept on file (and adapted, as required by the partner) for later years or in case an opportunity to work together develops in the future.
The word “Interdisciplinary” in “Master in Interdisciplinary Data Science” reflects the fact that data is collected in almost every domain, and data scientists are needed who can solve problems in all disciplines. MIDS embraces this idea and admits students who either want to apply data science to a topic they are already passionate about or have a strong background in or who demonstrate inherent curiosity for thoroughly understanding and solving new kinds of problems. By virtue of emphasizing background expertise, the majority of MIDS students have previous work or research experience, and some have extensive experience. In addition, the types of problems students are interested in addressing cover a variety of diverse fields, including medicine, finance, engineering, marketing, political science, and music. To learn more about this year’s students, see the MIDS student page.
Each student on the team is expected to spend approximately 15 hours a week on the project over the duration of the fall and spring semesters. Capstone partners should expect students to be available to work on the project throughout December and January (including the interlude between the academic semesters), but with an awareness that many students will take about a two-week vacation toward the end of December and the beginning of January, and that the students’ curricular course will not be in process during most of December and early January.
By the time MIDS students begin their Capstone projects, they will have completed a summer internship in data science and taken the following courses:
- Data to Decision
- Modeling and Representation of Data
- Principles of Machine Learning
- Data Wrangling and Principles of Text Analysis
- Data Management Systems
- Data Visualization, Logic, and Storytelling
- Data Science Ethics
- Data Science Dialogues Seminar
- Up to two additional electives in a technical area or subject domain
While they are completing the Capstone, they will take between three and six additional electives in a technical area or subject domain. Throughout both years of their coursework, they participate in workshops and continuous training in:
- Crafting appropriate metrics of success
- Stakeholder analysis and communication
- Project Management
- Team-building and team-participating
- Giving and receiving feedback
- Navigating disagreements with stakeholders and team members
- Design thinking
The Partner Liaison acts as the primary liaison between the partner organization and the project team and is actively engaged in the project. Partner Liaisons take the lead in shaping project direction – they articulate to the Capstone team the topic of interest to the partner organization and provide guidance to the team as it makes and implements its plans and reports its findings. Effective engagement requires that the Partner Liaisons hold regular meetings with the project team. This may involve visiting campus a few times during the semester to meet with the team, and/or communicating with the team through conference calls and the internet, using collaboration software such as Webex. Effective Partner Liaisons provide timely feedback to the team regarding the direction the project is evolving, ensure that the students can access organization data in a form that is appropriate to the project, and generally provide guidance with respect to organization-specific issues and policies. Ideally, the successful completion of the project is an objective recognized by senior management of the organization and a task for which the Partner Liaison is responsible to that organization.
While the amount of time will vary depending on the nature of the project, most Partner Liaisons will spend about 2 hours per week over the course of the project. Partner Liaisons are expected to be available for a mid-year review and to attend the final presentation either virtually or in person. Project teams make their final presentations at the end of April.
The 2-semester Capstone course devotes classroom time and has scheduled assignments dedicated to establishing team expectations, creating work plans with outside partners, providing weekly updates, performing 360-degree evaluations of team performance, and reflecting on management issues that arise. Students, faculty, and Partner Liaisons will establish specific commitments and work plans at the beginning of each project. These will vary across teams. In addition, trained project managers will support each project to ensure participating entities are engaged and communicate effectively.
Duke University can provide Capstone projects with a secure data environment and the software and computing resources required to complete the project. Capstone partners also have the option of requiring that students work in the partner’s computing space, in which case the partner is responsible for the associated costs. Students do not have travel funds, though they are welcome to accept travel support from Capstone partners.
Capstones and IGE Capstones have much in common – both involve vertically integrated research teams focused on providing deliverables to a partner. IGE Capstones, developed with support from the National Science Foundation, incorporate a Ph.D. student (known as an IGE Doctoral Fellow) into a MIDS Capstone project selected by the student and developed with that student’s involvement. IGE Doctoral Fellows participate in all Capstone classwork and activities (including meetings with stakeholders, teamwork and leadership workshops, analysis plans, milestone reports, and mid-year reviews) and have the opportunity to participate in any other professional development activities provided by the MIDS program (such as job fairs, interview preparation sessions, and seminars with non-academic parties who use data science), and will have the option to take MIDS core courses. IGE Capstones also receive $3,000 in research funds.
Duke Master in Interdisciplinary Data Science (MIDS) is a new curriculum that is in its first year – as such, no Capstone projects have been completed (or even started). However, Capstone projects will have much in common with Duke’s Data+ program that brings teams of students, post-docs, and faculty together with a partner during the summer and that, over 10 weeks, engages them in marshaling, analyzing, and visualizing data to answer a previously unaddressed question, to complete proof-of-principle work that may lead to subsequent additional collaborations, or to create tools that might facilitate community engagement with data and data-driven questions. Information about the 24 Data+ projects completed in the summer of 2018 can be found here.