Capstone Project (DSCI 8930)

About DSCI 8930

DSCI 8930 — MS Project is the capstone experience of the BDML concentration. It is a 1–4 credit hour course in which you demonstrate mastery of the full data science pipeline — from raw data to deployed, evaluated model — following an industry-standard methodology. It is the culminating academic deliverable of your BDML degree and serves as a tangible, portfolio-ready artifact you can present to employers.

Is the capstone required for graduation? Yes. DSCI 8930 is a degree requirement. You must complete it — and pass the oral final examination — to graduate from the BDML program.

What Is the Capstone Project?

The capstone is an independent data science project you design, execute, and present under the supervision of a BDML faculty advisor. It is not a course with weekly lectures — it is a project-based experience where you apply everything you have learned in the program to a problem of your choosing.

The project must follow one of the accepted industry-standard data science methodologies:

• CRISP-DM (Cross-Industry Standard Process for Data Mining) — the most widely used framework in industry, covering business understanding, data understanding, data preparation, modeling, evaluation, and deployment

• SEMMA (Sample, Explore, Modify, Model, Assess) — a methodology developed by SAS Institute emphasizing iterative model building and assessment

• KDD (Knowledge Discovery in Databases) — the foundational process model proposed by Fayyad, Piatetsky-Shapiro, and Smyth (1996), covering selection, preprocessing, transformation, data mining, and interpretation of patterns

You are not required to use a specific tool or programming language, but your methodology must be clearly documented and your results reproducible.

What You Will Demonstrate

A successful capstone project demonstrates all six core competencies of the BDML program:

1. Understanding the Principles of Data Science — framing the problem correctly and selecting appropriate methodologies
2. Data Collection and Visualization — acquiring, cleaning, and exploratively analyzing real-world data
3. Identification of Data Science Tasks — recognizing whether the problem calls for classification, regression, clustering, anomaly detection, or other approaches
4. Solution Design and Implementation — building and tuning models using appropriate algorithms
5. Analysis and Evaluation — rigorously comparing approaches using appropriate metrics, accounting for class imbalance and other real-world challenges
6. Interpretation of Outcomes — communicating findings in a clear, human-friendly way that translates technical results into actionable insights

The Oral Final Examination

Upon completion of your project, you will present your work in an oral final examination conducted by two BDML faculty members:

1. Your project advisor (Committee Chair) — the faculty member who supervised your project throughout the semester
2. A reader/reviewer — a second BDML faculty member who evaluates your work independently

The examination is not an adversarial defense — it is a structured conversation about your project, your methodology choices, your results, and what you would do differently. Strong preparation means knowing your project thoroughly and being able to speak confidently about every decision you made.

A note on standards: The requirements described on this page represent the minimum for a successful capstone. Many advisors and students set the bar considerably higher — producing work that becomes a genuine portfolio centerpiece for job applications, a research paper, or a foundation for PhD program applications. The capstone is one of the few opportunities in a taught master’s program to produce something truly your own. How much you invest in it is up to you — but those who treat it as more than a checkbox tend to leave the program with something they are genuinely proud of, and employers and admissions committees notice the difference.

Choosing a Topic

You have significant freedom in choosing your capstone topic. A good capstone project typically has the following characteristics:

• Real data — publicly available or provided by a partner organization, not synthetic
• Meaningful problem — something with genuine practical or research value, not a textbook exercise
• Appropriate scope — large enough to demonstrate the full pipeline, small enough to complete in one semester
• Personal connection — topics you genuinely care about tend to produce better projects and more confident oral examinations

Past BDML capstone projects have spanned domains including healthcare analytics, financial fraud detection, social media analysis, climate data, space weather forecasting, and sports analytics, among others.

Working With Your Advisor

You will need a BDML faculty member to serve as your project advisor. Reach out early — ideally the semester before you plan to enroll in DSCI 8930.

The best way to find an advisor is to approach an instructor from one of your BDML courses whose work genuinely interests you. You already have a personal relationship, they know your abilities, and they are familiar with BDML standards. This gives you the best chance of finding a high-quality advisor and a project that excites you.

When you reach out, identify yourself as a BDML student upfront. Many BDML courses also enroll MS-CS and MS-Applied Statistics students, so the instructor may not automatically know your program or its capstone requirements. Be explicit: tell them you are in the BDML concentration and that your capstone project must follow an approved data science methodology (CRISP-DM, SEMMA, or KDD). This matters because projects that are fine for MS-CS students — such as pure software engineering or systems development projects — are not eligible for BDML capstone credit.

How to propose a project: You have two good approaches:

• Bring your own data: Show the instructor a dataset you have access to and propose a project around it. Faculty appreciate students who come prepared — it signals initiative and makes the conversation concrete.
• Ask what they have in mind: Many faculty have ongoing research projects or datasets looking for a capable student. Simply asking “do you have something in mind that might fit a BDML capstone?” can open doors you did not know existed.

Either approach works. The goal is a project with real data, meaningful scope, and a clear methodology — and a faculty member who is genuinely engaged with what you are doing.

If you are unsure who to approach or need guidance, contact the BDML Program Director for suggestions.

Registration

Enroll in DSCI 8930 for 1–4 credit hours. The minimum required for graduation is 1 credit hour, which is what most students take. Additional credit hours are available for students who:

• Choose a more extensive or research-intensive project scope
• Spread the capstone work across two semesters

If you plan to enroll for more than 1 credit hour or across multiple semesters, discuss this with your advisor early — it affects your plan of study and DegreeWorks credit count. Confirm the appropriate enrollment with both your project advisor and academic advisor before registering.

GTA/GRA students: Check your appointment enrollment requirements before registering — your minimum credit hour obligations may affect how many hours of DSCI 8930 you should take in a given semester.

Important: Graduation Deadlines

The capstone process involves deadlines that are set not by the BDML program but by the Graduate School, the Registrar, and your academic advisor — and they are non-negotiable. Missing them can delay your graduation by a full semester.

At the very beginning of the semester you plan to graduate:

• Contact the academic program advisor (Ms. Naeemah Ahmed) to confirm your graduation application and degree audit are in order
• Discuss your intended defense date with your project advisor early — scheduling two faculty members for an oral examination takes time, and end-of-semester availability is limited
• Plan your defense date backward from the Graduate School’s submission deadlines, not forward from when you think you will be ready

Throughout the semester, stay on top of:

• Graduate School deadlines for degree completion and document submission
• Registrar deadlines for graduation application
• Any forms or paperwork your academic advisor requests

Practical advice: Set reminders for every deadline at the start of the semester. Do not assume your advisor, reader, or academic advisor will remind you — the responsibility for tracking these dates is yours. When in doubt, ask your academic advisor (Ms. Naeemah Ahmed) early rather than late.

Timeline Guidance

The Capstone as a Career Asset

Your capstone project is more than a graduation requirement — it is the most concrete evidence of your capabilities you will have when entering the job market. A well-executed capstone gives you:

• A GitHub-ready project you can share with recruiters
• A narrative for interviews — you built something real and can walk through every decision you made
• Confidence — you have solved a complete data science problem end-to-end, not just answered homework questions

Invest in it accordingly.

Contact

For questions about the capstone project, topic selection, or advisor assignments:

Dr. Rafal Angryk rangryk@gsu.edu BDML Program Director, Department of Computer Science, Georgia State University

BDML Quick Links

• BDML Program Overview — Program mission, value proposition, and contacts
• Already a GSU Undergraduate? — No-repeat rules and direct admit info
• 4+1 Dual Degree: BS-CS → BDML — Accelerated pathway for CS undergrads (Spring 2027)
• 4+1 Dual Degree: BS-DS → BDML — Accelerated pathway for DS undergrads (Fall 2026)
• Funding & Assistantships — GTA/GRA automatic consideration and requirements
• Data Science Internship — DSCI 8940 timing, requirements, and approval