Term Project

Design and implement a substantial big-data or privacy artifact, end to end

Project Manual

Overview

The term project for CSCI 49395Big Data Analytics and Applied Privacy (Fall 2026) requires you to design and implement a substantial, end-to-end big-data or privacy artifact and to report on it honestly. Your project must engage a real privacy or scale challenge (e.g. anonymization, differential privacy, a scalable pipeline, or privacy risk in ML) rather than simply calling a library. Separate what you measured from what you believe.

Deliverables

Submitted in order; each gates the next:

  1. Proposal — 1–2 page PDF: problem, dataset(s), method, privacy considerations, milestone plan. Must be approved before implementation.
  2. Code repository — Git repo with a README, a pinned environment, a fixed seed, and one-command reproduction. Raw data read-only; no committed secrets or PII.
  3. Final report (PDF) — 6–10 pages: intro, related work, method, results with effect sizes / uncertainty, a dedicated privacy analysis, limitations, conclusion.
  4. Jupyter notebook — runs top-to-bottom (Restart & Run All): intro → data-load → analysis → privacy analysis → conclusion.

Required Format

DeliverableFormatNotes
ProposalPDFTitle, names, references.
Code repoGit URLREADME + pinned env + seed; no secrets/PII.
Final reportPDFCaptioned figures; cite every dataset and claim.
Notebook.ipynb (v4)intro → data-load → analysis → privacy → conclusion.

Rubric

ComponentWeightWhat we look for
Proposal10%Clear question, feasible scope, privacy awareness.
Implementation40%Correctness, reproducibility, code quality.
Privacy analysis20%Explicit threat model + an evaluated mitigation.
Final report20%Honest claims, uncertainty, limitations.
Presentation10%Clear communication of problem and results.

Timeline

Dates follow the course schedule (Tue/Thu, 1:00 – 2:15 PM).

MilestoneDue
ProposalWeek 4
Progress check-in (repo + partial results)Week 11
Final report, notebook & presentationWeek 15

Examples & Templates

Get feedback from the instructor early and often. Never commit PII, credentials, or licensed data. Report the honest result, including nulls — proposals that need revision must be resubmitted and approved before you begin.

Submissions

Please log in to submit.