Assessment Plan¶
1. Institutional assessment scheme (confirmed)¶
| Institutional component | Marks |
|---|---|
| Sessional | 25 |
| Mid Exam | 35 |
| Final Exam | 40 |
| Total | 100 |
Only the three totals above are the institutional scheme. The internal composition of the 25-mark Sessional below is a course-level pedagogical choice — the instructor may adjust it, but the Sessional total of 25 marks is fixed:
| Sessional component | Count | Marks | When | CLOs |
|---|---|---|---|---|
| Labs | 32 | 8 (~0.25 each) | Weekly, due before the next session | 1, 2, 3 |
| Quizzes | 2 | 5 (2.5 each) | W8 S15, W15 S29 (session start, ~25 min) | 1, 2, 3 |
| Assignments | 2 | 6 (3 each) | See schedule | 1, 3 |
| Final project | 1 | 6 | Kickoff W12, presentations W16 | 1, 2, 3 |
| Sessional total | 25 |
- Mid Exam — 35 marks: W8 S16, 90 minutes; written portion + practical notebook task (CLO-1 with an entry-level CLO-2 task).
- Final Exam — 40 marks: W16 S32, 120 minutes; comprehensive written + practical (CLO-1, CLO-2, CLO-3, including a short AI-ethics / reproducibility section).
2. How your final grade is calculated¶
Final Percentage = Sessional + Mid Exam + Final Exam
Because the maximum is exactly 100 marks, the total is the percentage — no normalization is required.
Worked examples:
| Component | Max marks | Example 1 | Example 2 (perfect) |
|---|---|---|---|
| Sessional | 25 | 20 | 25 |
| Mid Exam | 35 | 28 | 35 |
| Final Exam | 40 | 33 | 40 |
| Total = Percentage | 100 | 81 | 100 |
| Letter grade | A- | A | |
| Grade points | 3.70 | 4.00 |
A ready-made Grade Calculator (Streamlit app) is included in the course repository so you never have to do this by hand:
streamlit run tools/grade_app.py
Enter your Sessional, Mid Exam and Final Exam marks (leave future components
empty); it shows the total, percentage, letter grade and grade points. The
same arithmetic is in tools/grade_calculator.py (pure Python, self-tested).
3. Official grading scale¶
| Grade | Percentage | Grade points |
|---|---|---|
| A | 85% and above | 4.00 |
| A- | 80–84% | 3.70 |
| B+ | 75–79% | 3.30 |
| B | 70–74% | 3.00 |
| B- | 65–69% | 2.70 |
| C+ | 61–64% | 2.30 |
| C | 58–60% | 2.00 |
| C- | 55–57% | 1.70 |
| D | 50–54% | 1.00 |
| F | below 50% | 0.00 |
4. Component details¶
Labs (8 marks of the Sessional)¶
Hands-on exercises completed largely in class and finished at home. Each lab has
stated objectives, a starter notebook, tasks, and 2–3 checkpoint questions.
Graded for correctness of the notebook, code quality, and answers to checkpoint
questions. See labs/.
Quizzes (5 marks of the Sessional — 2.5 each)¶
Short, closed-notes, in-class quizzes (multiple choice, code tracing,
debugging, short answer, and scenario items) covering the preceding weeks:
Quiz 1 = Weeks 1–7, Quiz 2 = Weeks 8–14. Full question banks with keys in
quizzes/ (quiz-1.md, quiz-2.md).
Assignments (6 marks of the Sessional — 3 each)¶
Larger individual tasks, submitted via a private GitHub repository (or as specified by the instructor):
- Assignment 1 — Data cleaning & EDA (CLO-1): clean a provided messy
dataset with a documented cleaning log, run a structured EDA, and write
evidence-backed findings. See
assignments/assignment-01-data-cleaning-eda/. - Assignment 2 — API + data acquisition + Streamlit app (CLO-1, CLO-3):
fetch data from an API/public dataset, clean and explore it, and ship an
interactive Streamlit application with user controls, documented in a README
and committed to GitHub. See
assignments/assignment-02-api-streamlit-app/.
Machine-learning modeling is assessed through Labs 17–22, Quiz 2, the final exam, and the final project's modeling component.
Final project (6 marks of the Sessional)¶
Team of 2–3 students; end-to-end data science project from a problem statement
to an analysis with a basic model and an AI-assisted or automated component.
Deliverables: proposal (W13), GitHub repository, presentation (W16), final
submission with a reproducibility + ethics reflection. See projects/.
Mid Exam (35 marks)¶
Written portion (concepts, code reading) + practical notebook task (CLO-1 with an entry-level CLO-2 task). Open-notes notebook allowed for the practical part as specified by the instructor.
Final Exam (40 marks)¶
Comprehensive: CLO-1, CLO-2, and CLO-3 concepts, including a short AI-ethics/reproducibility section.
5. Rubric sketch¶
Internal project rubric — the percentages below are applied to the project's 6-mark share of the Sessional; they are not course-level weights.
| Criterion | Weight (project) | Description |
|---|---|---|
| Problem framing & data acquisition | 20% | Clear question; appropriate, documented data sources |
| Cleaning & EDA | 25% | Sound cleaning steps; insightful exploration (CLO-1) |
| Modeling | 25% | Appropriate basic model; honest evaluation (CLO-2) |
| AI-assisted / automation component | 15% | Meaningful, working use of LLM/automation tools (CLO-3) |
| Reproducibility & ethics reflection | 10% | Environment, seeds, version control; bias/privacy considerations |
| Presentation & communication | 5% | Clear, well-paced presentation; answers questions |
Labs and assignments are graded on correctness, code quality, and completeness of required outputs; rubrics are included in each deliverable folder.
6. Policies¶
Submission¶
- All code-based deliverables are submitted via Git/GitHub (commit history visible). Notebooks must be executed and saved with visible outputs.
- Late submissions: −10% per day (or per program policy), up to 3 days; contact the instructor before the deadline for extensions (medical/family reasons).
Academic integrity & AI use¶
- Submitted code must be understood by the student; the instructor may ask any student to explain their work.
- Because CLO-3 explicitly covers AI-assisted workflows, AI tools may be used, with these rules:
- Disclose any AI assistance (tool, prompt, and how output was used) in a short note attached to the deliverable.
- Verify AI-produced code: it must run, and you must be able to explain it.
- AI assistance for exams and quizzes is not permitted unless stated otherwise.
- Plagiarism of peers' or external code without attribution follows standard institutional policy.
Attendance¶
2+ unexcused absences in the same module may require completion of a make-up task; per program policy.
Re-grading¶
Requests within 1 week of grade release, in writing, stating the specific issue.