Project Defense, Revision, and Final Submission

Project Defense, Revision, and Final Submission

The capstone defense is the professional moment your entire program has been building toward. You will stand before a panel of at least one instructor and one industry reviewer, present your end-to-end data science project in 20 minutes, answer 10 minutes of technical questions, and then deliver a polished final submission package. This lesson walks you through every stage: mock defense, formal defense, revision, and submission.

The Purpose of a Formal Defense

A project defense does more than show what you built — it demonstrates that you understand every decision you made. Industry reviewers ask hard questions precisely because your future employer will too. The defense format mirrors real workplace presentations: a data analyst who cannot explain why they chose a particular model or how they handled missing values loses credibility quickly. Think of the defense as a two-part test — the 20-minute presentation tests your ability to communicate findings, and the Q&A tests the depth of your understanding.

Stage 1: The Peer Mock Defense

Before the formal panel, you will run a full mock defense with two classmates acting as your panel. Use the official defense rubric — the same one the real panel will score you on — so feedback is specific and actionable.

Your peers will time you strictly to 20 minutes. If you run over, practice stopping at the mark anyway, because the formal panel will do exactly that. Peers score you on the five rubric dimensions: problem framing, data methodology, analytical rigor, communication clarity, and Q&A responses. They also provide written notes on chart readability, how clearly you explained your model, and how well you handled unexpected questions.

After the mock defense, you have 48 hours to revise your slides and talking points. Do not skip this window. Students who treat the mock defense as optional consistently struggle in the real panel. Common revision priorities include simplifying cluttered charts, sharpening the one-sentence summary of your key finding, and preparing crisp answers to questions your peers raised.

Stage 2: The Formal Defense

The formal defense follows a defined structure:

• Minutes 1–3: Problem statement — what business or research question did you address, and why does it matter?
• Minutes 4–8: Data methodology — where did your data come from, how did you clean and transform it, and what ethical considerations did you address (bias, privacy, data provenance)?
• Minutes 9–14: Analytical results — walk the panel through your Power BI dashboard live, highlight key visuals, and explain what your model produced and what its limitations are.
• Minutes 15–20: Recommendations and conclusions — what action should a decision-maker take based on your findings?

After your 20 minutes, the panel asks at least five technical questions. Questions may come from any of the following areas: your choice of data sources, how you handled outliers or missing data, why you selected one model over another, how you validated your results, and what ethical risks your project raises. Answer precisely and honestly. If you do not know something, say so and explain how you would find out — that response is almost always scored higher than a guess.

A note on your live dashboard: Test your Power BI Service link the morning of your defense on the room's internet connection. Have your .pbix file open locally as a backup. Technical failures are not an excuse for a poor defense, but a calm backup plan demonstrates professionalism.

Stage 3: Final Revisions

Within one week of your defense, incorporate all written panel feedback into your final report and dashboard. Read every comment carefully. Distinguish between suggestions (things you may choose to improve) and required corrections (errors in methodology or misrepresented data that must be fixed before submission). Revise your Power BI visuals if the panel noted readability issues, and update any section of your report where the Q&A revealed a gap in your explanation.

Stage 4: The Final Submission Package

Your complete submission must include all of the following, uploaded to the LMS by the posted deadline:

1. Cleaned data files — final, documented versions of every dataset used

2. Git repository link — your repository must have a clear commit history that shows iterative work, not a single upload

3. .pbix file — your Power BI Desktop file

4. Final report PDF — the revised, polished written report

5. Project reflection memo — one page describing what you would do differently if you were starting the project again

Missing any single file constitutes an incomplete. If your submission is incomplete, you have 10 business days to resubmit. Do not miss that window — it closes your opportunity to receive a passing grade for the capstone.

Scoring and Passing Standard

The defense rubric scores five dimensions on a 1–4 scale (total possible: 20 points). To pass, you must score at least 3 out of 4 on every dimension and reach a total score of at least 16 out of 20. A score of 2 on any single dimension, even with a high total, requires remediation. Review the rubric before every practice session so the criteria are not a surprise on defense day.

Professional and Ethical Responsibility

All data used in your project must be properly sourced and cited. Any dataset containing personally identifiable information must have been handled according to the data governance practices covered earlier in the program. During the defense, be transparent about your data's limitations — overstating the reliability of your results is an integrity issue, not just a technical error. Reviewers notice, and employers remember.