Project

Past project examples — looking for inspiration? Watch the YouTube playlist of past project presentations from previous years.

Proposal

Your proposal can be a "skeletal" version of your final report.  Here are some things you may wish to include.

1. Project Title and Team

• Provide a title for your project
• List all members of your team (full names and CNET IDs). The ideal number of group members is 2-3; please reach out to the staff if you need an exception on this and justify it in your proposal. In the above template, you can drop this in simply as the author list.

2. Project Summary

• Think of this as a draft of your project report introduction. Advice on writing a good introduction
• What is the problem?
• Why is it an important problem? (Provide citations if possible to support your claims of relevance. Examples could be previous work/research papers, news stories, anecdotes, software packages, etc.)
• What kinds of actions could be taken based on your results, and who is equipped to take those actions? 

3. Related Work

• Draft as much as you can of your related work section.
• Organize your related work so it is more than just a “list”.  Organize thematically; look at related work sections from other published papers for examples.
• Include all references (e.g., as bibtex) so your related work is as complete as possible for now.
• Also articulate how your work extends the related work you cite.

4. Project Preparation and Prerequisites

• What do you need to complete this project?
• If your project is measurement based, describe the data you have and the data you’ll need to collect, as well as results you might expect and how you would present them
• If your project is to build a system (e.g., censorship circumvention tool), describe how you may go about doing that (will you extend an existing software code base, for example)?

5. Evaluation

• This of this as a starting point for the project.
• Describe your evaluation process? Be as specific as possible about your evaluation techniques 

6. Ethics

• Briefly discuss the ethical implications of your work. Will your models suffer from any kind of bias? Will this potential bias impact the fairness of any proposed solutions?

Final Deliverable 

Your final project deliverable should be:

• A brief (4-6 page is plenty) writeup of your project. building on your proposal/interim report in the ACM paper format we discussed in earlier assignments (you should have a nice head start on this!).  There is no page limit (or minimum) for this writeup.  It can be "as long as it needs to be".  If you did some data analysis, the writeup should obviously make it clear exactly how you collected, processed, and analyzed the data to a level that a reader should understand what you did, and so forth.  
• A brief video summarizing your project and main findings. This can be (and should be!) short!  Examples from past years are available in this YouTube playlist.  I intend to use these to (1) Put them into a playlist and host a watch party (if/when there is interest); (2) showcase the class; (3) show them as examples in future years.

 Use this (preferred) or this (if you must) as a paper skeleton for your report.

Grading (100 points)

The project is 35% of the course grade and is scored out of 100 points against the brief above: 15 for the proposal, 85 for the final deliverable. Points reward understanding and an honest, complete account of what you did, not polish. The writeup is graded from its text, so put the numbers, figures, and citations in the paper itself rather than pointing at a notebook.

Part 1: Proposal (15 points, graded for completion)

A complete, good-faith proposal earns all 15 points. The proposal exists so the instructor can catch scope, data, and ethics problems early, not to judge the idea. Vague sections get a comment and a request to firm them up, not a deduction, unless the section is missing. One proposal per team; every member should have a copy (or a pointer to it) in their own repo.

#	Section	Points	What counts as present
1	Title and team	1	A title; every member’s full name (the author list); team of 2–3, or a justification for a different size.
2	Project summary	3	States the problem, why it matters (with at least one citation: paper, news story, software, or anecdote), and what action could follow from the results and who could take it.
3	Related work	3	Organized thematically rather than as a list; references included (e.g. BibTeX); says how your work extends what you cite.
4	Preparation and prerequisites	3	Measurement projects: the data you have, the data you still need, the results you expect, and how you would present them. Systems projects: what you will build and which existing code base, if any, you will extend.
5	Evaluation	3	Names specific evaluation techniques or metrics, not just “we will evaluate it.”
6	Ethics	2	At least a paragraph on the ethical implications of the work, possible bias, and whether that bias affects the fairness of any proposed solution.

The interim check-in (how it’s going, obstacles, an outline, a meeting in office hours) receives feedback, not points.

Part 2: Final deliverable (85 points)

Component	Points	What earns full marks
Introduction	10	States the problem in the first page; explains why it matters with citations; names who can act on the results and what they could do.
Related work	10	Thematic subsections or paragraphs, not a list; each theme states how your work differs from or extends it; every in-text citation resolves to a complete bibliography entry.
Method / Data	15	Measurement projects: what data, how it was collected (dates, vantage points, tools, sample sizes), how it was processed and filtered, and how it was analyzed, at a level a reader could reproduce. Systems projects: what was built, its architecture, which existing code base was extended, and how it was tested. Code and data are released, or the paper says why not.
Results / Evaluation	15	The evaluation matches what the proposal promised, or the paper explains the change. Results are specific: figures and tables with labelled axes, captions, and the actual numbers pasted in. Each claim in the text points to a figure, table, or number. Null or negative results are reported as such.
Ethics and limitations	8	Ethical implications specific to your data or system (consent, risk to users in censored regions, bias), plus concrete limitations of the method and data and what would be needed to strengthen the claim.
Clarity, formatting, team roles	7	ACM format via the paper skeleton or the Overleaf template; roughly 4–6 pages (longer only if it needs to be); abstract; numbered sections; author list; a short statement of who did what.
Video	10	A short video (a few minutes; link or file in the repo) that states the question, the method, and the main finding, and is understandable without reading the paper.
Reflection & AI-verification	10	A “Reflection” section or appendix that reports (a) at least one thing you tried that did not work and what you did instead, (b) what surprised you in your own data or system, and (c) if you used an LLM, at least one claim you checked against your data, a primary source, or a running system, and what you found. Include an AI-usage appendix (prompts, how you verified the output) if you used one. Generic prose that could describe any project earns little.
Extra credit: submission-ready artifact	+10	Either a public dataset or code release with a README that explains how to reproduce the paper’s figures, or a writeup that is ready to submit to a named workshop (its format, its page limit, complete).

Requirements that gate the score

• Deadlines. The proposal and the final paper and video are due on the dates posted on the course site; the syllabus late policy (96-hour grace balance for the term) applies.
• Evidence in the text. Cite figures, tables, and numbers so a grader can find your evidence in the paper. A correct, specific two-sentence answer beats a page of vague prose.
• Per-team variation is expected. Teams choose different questions, data, and systems. You are graded on whether the analysis is correct for what you chose, not against a fixed answer.
• AI use. Encouraged, with acknowledgment and an explanation of how you prompted and verified it (see the Reflection item).
• Defendable live. Any team member may be asked to explain or reproduce any part of the project in office hours, in class, or during the presentation, per the syllabus.