DATA 750 – MSDS Thesis
- All MSDS students must complete a thesis (DATA 750) as part of their graduation requirements.
- Typically, the thesis is completed in the last semester of the student’s degree program.
- Prior to registration, students must have an approved proposal for their project. The proposal must have signatures from the thesis advisor and two readers.
- In order for a proposal to be approved prior to registration, the student must have started a dialog between the advisor and readers at least 4 weeks prior to registration. It is recommended that students initially contact (select) their advisor during the early part of the preceding semester to begin the discussion.
- Students do not need fully elaborated project plans to start the dialog with their advisor – the advisor will assist in solidifying and clarifying the project. Students do need to have an idea, topic, and rough goal to start the discussion.
- Projects that require the collection or evaluation of personal information may require approval from the College’s Institutional Review Board (IRB). The thesis advisor will help determine whether the project requires IRB approval – which must be applied for prior to DATA 750 registration. Students who plan to use data sets that include personally identifiable information or actively collect information from individuals must begin discussing their project towards the beginning of the semester preceding their thesis, as IRB approval can take several weeks.
A Master’s level thesis project in Data Science can take on many different shapes, and students are encouraged to begin discussing their ideas with potential thesis advisors and readers as soon as possible in order to mold their project ideas into a suitable thesis.
Thesis projects may be characterized as (but are not limited to) any of the following:
- Technique-focused: Improving upon existing, or developing new mathematical and computational techniques for a specific Data Science problem.
- Tool-focused: The creation of a computational tool that can be used to solve or present specific types of generalizable Data Science problems.
- Domain-focused: The application or development of a Data Science tool or technique to a specific domain to answer domain-specific questions relevant to the stakeholders within that domain.
The key commonality among all Master’s theses is that students must demonstrate the value of their project’s outcomes. The thesis must contribute useful knowledge to the Data Science community or a particular domain/industry. While the deliverable of a project may be a technique/tool to solve specific or general categories of Data Science problems, projects may also produce visualizations, dashboards, or meta-analyses that expand upon and synthesize existing literature and data to build on knowledge within the field and/or explore ethical implications of data.
The project’s scope must be realistic, to fit within the constraints of a 15-week semester – while also substantial enough to warrant an entire semester of work (DATA 750 is a 3-credit course). Just
A necessary component of any thesis is some degree of risk. When proposing a project, the student advisor will explore what aspects of the project may prove more difficult than expected.
- When a project involves using data sets to answer new questions using new techniques, the answers to those questions may not be what was expected.
- When developing a new mathematical or computational technique to better solve a Data Science problem, it may turn out that it does not!
The evaluation of your thesis centers around the process in which you went about completing your project – not solely on the outcome/deliverable. Your thesis committee will work with you throughout the project to guide you through the expected challenges you encounter – which is why constant communication with your thesis committee is so crucial to your success.
DATA 750 – Thesis in Data Science – Proposal (DOC)
Student Thesis Handbook (PDF)