The project proposal consists of three parts:

  1. Writing a carefully stated and motivated proposal that the group will follow in Quarter 2.
  2. Laying out an initial, detailed plan for executing on the proposal.
  3. Presenting a short ‘elevator pitch’ for the proposal.

This assignment will be worked on in the same Capstone Project group that will work on the project through Quarter 2.

Considerations when choosing a problem to investigate

This proposal will form a contract for what will be worked on throughout the rest of the Capstone Sequence. Tou will have to work together and make decisions as a group, even if you may not 100% agree with the choices. As you develop your proposal consider the following observations:

  • Try to agree on a general theme for what you potentially want to investigate in your project.
  • You do not need to have 100% agreement on the proposal, rather you need to be comfortable that a general area of investigation is interesting to you. The specifics of proposals change as you work on it, and you may have to change direction anyway.
  • Discuss how your strengths might complement a teammate’s strengths. A proposal can consist of quasi-independent pieces, that fit individual team-member’s strengths (e.g. Buidling and designing an ML model, a statistical investigation into data generated by a learning model, visualizations for results). All of these would be too much for a single person in a quarter; a team could manage it though.

Proposal

The proposal will form a plan for your group’s Quarter 2 project. The proposal will consist of the following parts:

  • A description of the problem, along with its similarity and differences from the problem being studied in the current domain. In particular, it should note the scope of the data being considered for the investigation.
    • Give the reader enough context to convince them the problem is interesting and worth looking into.
    • Provide enough detail for a domain expert to understand whether the proposal is reasonable and realistic.
  • Specify the project output (a report/paper, website, a product that does something).

Note: if you are incorporating data into your project that you have not yet used, you must get access to it in Q2 and show (via a data analysis) that it contains what you need to do the project.

Schedule

The proposed schedule should include:

  • A proposed 6 week schedule with weekly goals and tasks for executing the proposal.
  • A delineation of responsibilities among group members in the 6 week schedule.

Note: the schedule is 6 weeks to allow for tidying-up details, work on effective communication, and dealing with inevitable set-backs.

Elevator Pitch

You will need to write and present an Elevator Pitch for your proposal. The elevator pitch should be under three minutes and summarize the project proposal, and its relevance, to a non-specialist. The elevator pitch will be:

  1. Recorded (as an mp4) and turned in on the due date specified on Canvas.
  2. Presented orally to the domain discussion group (either in section, in lab-hours, or on video, depending on the preference of the domain).

Proposal Content Checklist

A thorough proposal should answer the following questions:

  • What set of problems are being investigated? (Or what are you proposing to build, and what problem does that solve)?
  • How do these problems relate to the domain Q1 project?
    • Has previous work attempted to answer these questions?
    • If not, why are these questions considered interesting? If so, how did they fail?
    • In what way does your investigation address a deficiency in the replication project?
  • What are the specific methods that will be used to attempt to solve (a subset of) the problems in your proposal? Be specific about the data and techniques you plan to use.
  • What is the project output? Is it a paper, website, work of art, tool? If your project analyzes data, what ways do you plan on communicating the analysis? If your project generates data, you still need an analysis of the generated data it produces!
  • For the proposed schedule, look to the domain schedule (result replication) as a guide.