This page is your one-stop map of capstone deliverables, process expectations, and milestone deadlines. It is more descriptive than the operational DS3 framework pages on the course site, but it is still a summary. When anything here disagrees with a dated Canvas assignment or the syllabus, follow Canvas and the syllabus.

Deep operational reference (roles, boards, Brief/Critique mechanics).

Per-milestone rubrics and submission checklists: links in each milestone section below and on the matching Canvas assignment.

The DATA 510 capstone is a semester-long culminating project: you propose, build, evaluate, and communicate a consequential data science outcome you can defend to peers, faculty, and industry guests at the end-of-term poster session. It should read as one coherent story, not a pile of disconnected homework.

A strong capstone integrates five pillars from your MS coursework. Your proposal and every milestone should make the integration explicit:

• Data engineering: sourcing, cleaning, storing, versioning, and automating data so analyses are reproducible and auditable.
• Analytics and machine learning: appropriate modeling or inference, validation discipline, and honest limits of what the data support.
• Visualization and communication: graphics and narrative legible to technical and non-technical audiences (poster, write-up, website).
• Ethics and responsible use: privacy, consent, fairness, transparency, and proportionality in collection, sharing, and interpretation.
• Statistics and research design: clear questions, baselines, and evidence that match the claim you want to make.

Projects that read as “only a dashboard” or “only a notebook” without engineering, ethics, or communication will be pushed to deepen before approval. Data sources must be selected and approved by the instructor early in the term (request approval in your project’s #blockers Discord channel).

Owner team: Solo, or a self-selected team of two or three students. Teams of 2–3 must carry noticeably higher scope than a comparable solo project (more integration, evaluation, documentation). Team scope is reviewed and approved at the Project Proposal (M1).

Your studio: Your owner project plus two or three peer Stakeholder Product Owners (peer POs) assigned from adjacent capstones. Peer POs review your work weekly through written Studio Briefs and Studio Critiques. You are not responsible for their project deliverables; they are a standing review panel, not extra teammates.

Dual-role rule: Every student is simultaneously:

• Owner on exactly one project (your capstone), and
• Peer Stakeholder PO on exactly two adjacent projects (one in each Studio Session other than your own).

Find your peer PO assignments and Studio Session (1, 2, or 3) on the Peer Stakeholder PO lookup.

Owner Product Lead: One named student per project leads the weekly Studio Session, maintains the backlog and board, and presents the Iteration Review. Solo students are their own lead.

Component	Weight	What it covers
Professional development activities	10%	Structured career prep on alternating weeks (networking, interviews, portfolio).
Weekly peer and meta-project evaluations	5%	Short reflections on your team and the two peer projects in your cluster.
Project process and practices	5%	Git workflow, chartering, weekly summaries, commit quality (especially last five weeks).
Technical writing practice	5%	In-class writing exercises (partial credit only for makeup if you miss class).
Stakeholder participation (peer PO duties)	15%	Written Studio Brief and Studio Critique each week for each of your two peer projects, plus live Studio Session participation. Missing 4+ Briefs or Critiques caps this component at 50%.
Project proposal (M1)	9%	Week 4 roadmap: question, data, methods, ethics, DS3 tooling.
Data summary (M2)	9%	Week 7: data gathered or pipelines stable; ingestion frozen after this.
Poster rough draft (M3)	8%	Week 10: full end-to-end poster story (draft figures OK).
Write-up rough draft (M4)	9%	Week 12: publication-style report; methods and findings freeze after this.
Final poster presentation (M5)	18%	Week 14 live session + final poster file; no grace period for the session.
Final write-up (M5)	14%	Polished permanent record; due after poster session (see Canvas).
Project website and dissemination	8%	Public GitHub Pages (or equivalent) linking write-up, poster, repo.

Starting week 4, the course runs on Data Science Studio Scrum (DS3). DS3 is how you steer week to week toward graded milestones; it does not replace them. Think of it as the operating system for your capstone.

Before each class (owner team):

• Post an Iteration Review section in your repo README.md: completed PBIs, in-flight work, observations, analysis decisions, plan for next week.
• Execute on pulled backlog items through the board columns Create → Observe → Analyze → Done.
• Push Git commits often; unpushed work is not backed up.

Before each class (peer PO, for each of your two peer projects):

• Commit a written Studio Brief to studio/briefs/ (requirement, question, or risk for the owner team’s next iteration).
• Commit a written Studio Critique to studio/critiques/ (assessment of last week’s delivery against the charter and prior brief).
• Drop links in the project’s #studio Discord channel.

In class (Block 2, three 40-minute Studio Sessions): Owner standup, Critique discussion, Brief discussion, backlog refinement and board walk, instructor sync. Full agenda: Studio Session weekly ritual.

Good peer feedback: ties comments to visible evidence (board, README, plots); separates curiosity from blocking risks; surfaces ethics and engineering concerns early.

Each owner team maintains one GitHub repo created from the DS3 template:

• Template: github.com/LucasCordova/ds3template
• Projects board template: copy LucasCordova/projects/3 (columns: Backlog, Create, Observe, Analyze, Done)
• Discord: class server; opt into your project category and both peer project categories

Top-level repo layout:

README.md          # weekly Iteration Reviews
CHARTER.md         # vision, mission, SLAs, DoR/DoD
BACKLOG.md         # human-readable backlog mirror
studio/briefs/     # peer PO Briefs (written, committed)
studio/critiques/  # peer PO Critiques (written, committed)
src/               # production code
notebooks/         # exploration and analysis
data/              # raw external/interim gitignored; processed committable
deliverables/      # milestone PDFs and source (see below)

Collaborators required: LucasCordova (Admin on repo and board); each peer PO (Triage or higher on repo; Read or Write on board per charter).

Milestone folders under deliverables/:

• M1-proposal/   M2-data-summary/   M3-poster-draft/   M4-writeup-draft/   M5-final/

Quarto encouraged for write-ups (commit .qmd source plus rendered PDF/HTML). Word, Google Docs, LaTeX, and other editors are permitted; graded milestones still need a PDF in the correct folder. PDFs cannot be line-merged in Git; teams must coordinate who renders and commits milestone PDFs.

Every backlog item should carry a milestone tag (M1-proposal, M2-data-summary, M3-poster-draft, M4-writeup-draft, M5-final, plus infra or ethics as needed). Tags without a milestone are a smell.

Week	Milestone	Weight	What it proves
3	Studio Charter (repo URL on Canvas)	—	Committed CHARTER.md, seeded backlog, live board, collaborators, Discord.
4	Project Proposal (M1)	9%	Locks direction, data plan, methods, ethics, team scope, DS3 tooling. Default roadmap for the term.
7	Data Summary (M2)	9%	Data gathered and documented, or pipelines run without babysitting. Ingestion freezes after this.
10	Poster Rough Draft (M3)	8%	Complete end-to-end poster story; placeholder figures OK.
12	Write-Up Rough Draft (M4)	9%	Publication-style prose across all report sections. Methods and findings freeze after this.
14	Final Poster Presentation (M5)	18%	Live conference-style session (both cohorts); final poster PDF. No grace period for the session.
14+	Final Write-Up (M5)	14%	Polished permanent record; due after poster session (Canvas lists exact date, typically 8/17).
term	Project Website	8%	Public showcase (GitHub Pages): links to write-up, poster, repo, reproducibility.

Also due most weeks: weekly summary report on Canvas; technical writing outputs on assigned weeks; Git and stakeholder Brief/Critique commits in repo.

Poster and write-up milestones share the same substantive components. Draft milestones reward completeness and honest work-in-progress; final milestones reward polish and accuracy.

Every poster and write-up must address:

1. Title, team, and affiliations
2. Abstract (concise hook plus takeaway on poster; publishable summary on write-up)
3. Introduction: problem and research question(s)
4. Impact and stakeholders
5. Related work or background (brief, cited; write-up; optional condensed on poster draft)
6. Data and data engineering (sources, scale, pipeline, reproducibility)
7. Ethics and responsible use
8. Methods and evaluation (statistics, ML, validation, baselines)
9. Results and visualizations (labeled figures; claims match evidence)
10. Discussion, limitations, and future work
11. Conclusions and recommendations
12. Reproducibility and references (repo URL, dependencies, citations; QR code optional on poster)

Poster format: Conference layout (often 48×36 in or A0); readable at arm’s length; high-contrast fonts; minimal dense paragraphs. Write-up length: typically 8–15 pages equivalent at 11–12 pt (figures and references count toward clarity, not padding).

Freeze rules (important):

• After M2 Data Summary: no further active ingestion work.
• After M4 Write-Up Rough Draft: no substantive methodology or findings changes (editing and polish continue).
• Final write-up must match what you defended at the poster session.

Quarto skeleton (optional starting point for M4): writeup-skeleton.qmd on the course site.

One PDF (~4–6 pages). Extends your committed CHARTER.md into a full roadmap. Required sections:

1. Title and team metadata (Owner Product Lead, peer POs, Studio Session, solo vs team)
2. Problem framing and research question(s)
3. Impact and stakeholders (including Shared Theme from peer PO lookup)
4. Data sources and access plan (instructor approval status for each source)
5. Data engineering plan (repo layout, pipelines, raw vs processed layers)
6. Planned analysis (methods, evaluation, Create/Observe/Analyze mapping)
7. Program integration (five pillars)
8. Ethics and risk notes
9. Timeline through M5 with explicit M2–M5 markers and risks
10. DS3 process and tooling (repo URL, Projects board URL, Discord, charter confirmation, Brief/Critique plan)

One PDF pointing to repo paths wherever possible. Required sections:

1. Research question reminder (links frozen data scope to the proposal)
2. Data inventory (source, license, size, date range, refresh policy, repo location)
3. Schema and organization (keys, joins, raw vs analysis-ready layers)
4. Ingestion status (manual complete or automated pipeline; evidence it runs reliably)
5. Data quality and profiling (missingness, duplicates, outliers; mitigations documented)
6. Reproducibility (commands to rebuild analysis-ready data; environment notes)
7. Ethics and access controls (PII, consent, sharing constraints)
8. Freeze statement: explicit confirmation that no further ingestion work is planned

Conference-style poster session in Salem on the last class evening, joint across both cohorts, with faculty and industry representatives.

You must:

• Submit final poster PDF per Canvas instructions
• Print or display as event logistics require
• Attend the full poster block with all team members present at the board
• Deliver a 60–90 second elevator pitch and answer depth questions without reading the poster verbatim
• Explain methods, validation, and limitations to technical and non-technical visitors
• Visit peer posters when not presenting

Public URL (typically GitHub Pages) suitable for employers. Minimum content:

• Project landing page (title, team, summary, hero visual)
• Links to final write-up (PDF and/or HTML), poster PDF, and GitHub repo
• Results highlight (key figures or short demo with captions)
• Reproducibility (clone, install, rerun main results)
• Ethics and data notice (licenses, attributions, use constraints)
• About the team; linked from your portfolio or GitHub profile

• Canvas: upload milestone PDFs (or submit repo URL where the assignment specifies). One submission per team unless the assignment says otherwise; list all members.
• GitHub: commit milestone source and rendered PDF/HTML under deliverables/; peer POs and the instructor should open artifacts without re-rendering.
• Late policy: 24-hour grace window on most deliverables, then 20% per day. Exceptions (no grace): final poster session, technical writing assignments, Git-related assignments as labeled on Canvas.
• Iteration expected: proposals evolve when data talk back; document changes in Iteration Reviews and charter amendments.
• Reproducibility: a peer should reorient mid-semester from README.md and CHARTER.md alone.