The specifications of the Quarter 1 Project have been finalized for Fall 2023.
As you know, your main deliverable throughout the capstone sequence is your Quarter 2 Project. In order to ensure that you’ll be able to successfully execute your project, you’re required to submit a detailed proposal.
The proposal consists of three parts. The proposal is worth 15% of your DSC 180A grade; each component of the proposal is worth the same amount (5% of your DSC 180A grade). Below, we describe each component and its deliverables in more detail. All components are due on Monday, December 11th at 11:59PM to Gradescope.
Quarter 2 Projects must be worked on in groups of 2-4; you cannot work individually. You must submit your proposal with the same group that you’ll work on the project with. It is up to your mentor to determine how groups are formed; you should ask them how groups will be formed. You’re required to sign up and attend the Week 9/10 TA Check-In in groups, so you should try and identify your group sooner rather than later.
After this year, whether you work in industry or academia, you will often have to write project proposals.
- As a data scientist in industry, you’ll often have to write proposals to justify why your idea has value and is worth investigating.
- In academia, to receive any sort of funding, you’ll need to write a grant proposal, which reads similarly to a project proposal.
As such, you are required to write a project proposal here, not just to better plan your Quarter 2 Project, but to gain this valuable skill. You must write a proposal even if your mentor has already determined what your Quarter 2 Project is going to be.
As you start to think about your Quarter 2 work, it’s a good idea to look at what students in your shoes last year were able to accomplish in the span of 10 weeks. Below, you’ll find some selected examples, though you can see all of last year’s projects at dsc-capstone.org/showcase-23, and all capstone projects ever at this sheet. You’ll learn more about the specific deliverables you’ll be required to produce in Quarter 2 in January (though you can read a draft of the full project spec here if you’re curious); for now, just focus on the big picture.
- BlockBazaar - Smart Contracts for Decentralized E-Commerce
Matin Ghaffari, Yu Huang, Alan Amirian, Wenyuan Chen (Mentor: Sheffield Nolan from Franklin Templeton)
🪧 Poster • 🌐 Website • 📖 Report • 💻 Code
- FinDL: Deep Learning Library for Finance Applications
Nathan Ng, Richard Tang, Gao Mo (Mentor: Zhiting Hu)
🪧 Poster • 🌐 Website • 📖 Report • 💻 Code
- Evaluating Fungal Feature Importance in Predicting Life Expectancy for Cancer Patients
Ethan Chan, Mark Zheng, Benjamin Sacks (Mentor: Rob Knight)
🪧 Poster • 🌐 Website • 📖 Report • 💻 Code
- Written Proposal (Graded by your mentor)
- Schedule (Graded by your mentor)
- Elevator Pitch (Graded by your TA)
- Summary of Deliverables
The most important component of your proposal submission is a written proposal. Your written proposal should:
- State the problem being investigated, for both a general audience who knows nothing about your domain and a domain expert (like your mentor).
- Give the reader enough context to convince them the problem is interesting and worth investigating – if the project is executed, what will its impact be?
- Provide enough detail for a domain expert to understand whether the project’s goals are realistic to complete using the methods described and in the timeframe allocated.
- Specify the project’s primary output, e.g. a report/paper, a website, or an application. (Note that even if your primary output is a website, you will still have to write a paper, and vice versa.)
Like the Introduction section of a paper, which we discussed in Lesson 4, a written proposal should consist of three parts, that start broad and get increasingly narrower. In fact, you will reuse a lot of your written proposal when writing your Quarter 2 Project’s introduction section – and you may even see a lot of overlap between your written proposal and Quarter 1 Project report.
Specifically, your written proposal should have three sections:
A broad problem statement, with context to justify the investment in spending 10 weeks on the project. This should be accessible to anyone – for instance, methodology staff, your family and friends, and hiring managers for the jobs you apply to should all be able to read this and understand what you are proposing and why it matters.
A narrow, careful problem statement for the domain expert. This section should discuss how these problems relate to the Quarter 1 Project in your domain. Has previous work attempted to answer these specific problems? If so, how did they fail? If not, why are these problems interesting? In what way does your investigation address a deficiency in the Quarter 1 Project? This section should be quite technical – make sure to refer to specific methods.
A statement of the primary output (i.e., state whether you will create a report/paper, a website, or an application). If your project analyzes data, specify how you will communicate the analyses. If your project generates data, how will you analyze the data it produces?
In addition to the above structure, you need to justify why your project will be successful. In particular, if your project relies on data that was not used in the Quarter 1 Project, you need to prove the following:
- You can obtain the data.
- The data contains the information needed.
- The data is of sufficient quality.
The easiest way to do this is to actually obtain the data while writing the proposal and perform some preliminary analyses to ensure it will actually be able to help you accomplish your goals.
In years past, many groups fell in the trap of waiting until Quarter 2 to obtain their data, and by the time they realized that they either can’t obtain their data or that it won’t actually help them with their project, they had wasted a significant portion of the quarter. Don’t let this be you!
Your proposal will form a contract for what you will work on in Quarter 2. You will have to work together and make decisions as a group, even if you may not fully agree with the choices. As you develop your proposal, consider the following:
- Try to agree on a general theme for what you want to investigate in your project.
- You do not need to have 100% agreement on the proposal, rather you need to be sure that the proposed area of investigation is interesting to you. The specifics of a proposal change as you work on it, and you may have to change direction while working on the project itself anyway.
- Discuss how your strengths might complement a teammate’s strengths. A proposal can consist of semi-independent pieces that fit individual team-member’s strengths – for instance, one student can be responsible for building and designing an ML model, another can be responsible for a statistical investigation into the data generated by the ML model, and another can be responsible for developing custom visualizations of the results. All of these would be too much for a single person to do well in a quarter, but they are all achievable when working in a group.
Just as you’re required to with your Quarter 1 Project, you’ll be required to submit a contributions statement along with your Quarter 2 Project at the end of Quarter 2. Last year, we saw several cases where a student in a group contributed a single-digit percentage of the overall work done by the project group, and those students’ scores suffered. One of the main purposes of the proposal is to avoid instances of this!
As of the release of these specifications (the end of Week 7), you may not be sure what to propose for Quarter 2. That’s totally fine. To help you get started, part of Methodology Assignment 5 will have you brainstorm ideas based on your current understanding of your domain.
Note that the amount of freedom you have in proposing your own project will depend on your mentor. In fact, your mentor may already have determined what your Quarter 2 Project will be, and that’s fine. In any case, your mentor will play a crucial role in your proposal, as they’ll be able to help you come up with a plan that you can realistically execute in 10 weeks under their guidance. Your final proposal submission should not surprise them at any point, because you should be discussing it with them regularly and incorporating their feedback.
You’ll submit your written proposal as a PDF to Gradescope as a group. Like with your Quarter 1 Project, you must use our LaTeX template on Overleaf. Your mentor will grade your written proposal according to this rubric.
You’ll need to create a detailed plan of what you will do each week in order to execute your proposal and complete your Quarter 2 Project. Your proposed schedule should include:
- A 6 week schedule with tasks and goals to be accomplished each week.
- A delineation of responsibilities among group members in the 6 week schedule.
By writing your schedule now, it’ll make it clear to both your group and your mentor whether your project can be realistically completed in the given timeframe. When creating your schedule, use your domain’s schedule and pace in Quarter 1 as a guide. What did you do each week? How much were you able to accomplish each week?
- Your schedules won’t be set in stone, as you will encounter hiccups along the way. However, by having a “contract” of who is going to do what, when, you’ll be able to hold everyone in your group accountable and will be able to help them if and when they get stuck.
- The amount of work you plan to do should scale with the size of your group. Groups of 4 should expect to accomplish more than a group of 2. It will be clear from your schedule whether your plan is too ambitious or not ambitious enough.
- DSC 180B – the capstone course you’ll be enrolled in next quarter – is a 4-unit course, meaning that you should expect to spend 12 hours per week on it on average. Plan for 11 of those hours to be spent on your project. Quality work takes time!
Why a 6 week schedule in a 10 week quarter?
As mentioned above, not everything will go according to plan and you will face setbacks. By planning to take 6 weeks instead of 10, you’ll have some breathing room to deal with these setbacks.
- Once you complete your tasks, you’ll need to spend a lot of time polishing your work and making it presentable to a broad audience. Remember, you want to build something you’ll look back on in years from now and be proud of, not something you just threw together at the last minute.
- In the rare case that you finish everything you wanted to before the deadline, you’ll surely be able to pose new questions and add more tasks.
How specific should the tasks be?
As specific as possible.
- Others in your group (and your mentor) should be able to tell exactly what you’re working on by reading a task description in your proposal.
- The more specific the description, the better. One rule of thumb is that you should never have the same task listed across multiple weeks – if you do, the task is not specific enough! Don’t just say “EDA.”
How should we divide tasks up among group members?
Divide tasks so that they take advantage of the strengths of your group members and so that they can be parallelized. Don’t divide tasks so far that you have no idea what your partners are working on and are unable to help them!
You’ll submit your schedule as a PDF to Gradescope as a group. It will be separate from your written proposal PDF submission. We don’t have any special formatting instructions for your schedule, though one idea is to make a table in a Google doc with one row per week and one column per group member, and submit a PDF of the doc. Your mentor will grade your schedule according to this rubric.
Finally, you’ll need to write and present an elevator pitch for your proposal. Your elevator pitch must be between 1 and 3 minutes long and should summarize the written proposal and its relevance for an audience that is not familiar with your domain. For instance, suppose you have a job interview in April and someone says “Hey, I saw you did a capstone project. Tell me about it!” You’d give them your elevator pitch.
Elevator pitches are purely verbal – that is, they should not have any accompanying graphics. They should include few implementation details, and instead focus on the problem statement, context, and motivation.
The typical anatomy of a short presentation is as follows:
- Context: What’s the background? Explain why this area is interesting.
- Complication: Tell us what solutions currently exist to the problem at hand, and where the “gaps” in current solutions are.
- Question: Clearly state the question that addresses the problem.
- Hypothesis: State your proposed solution. Propose a clear, quantifiable hypothesis that will advance our understanding.
This framework should sound familiar, and that’s because we discussed something similar in Lesson 4. You may want to re-watch Professor Benjamin Smarr’s video as you start to write your elevator pitch.
Unlike the other components of the proposal, pitches will be delivered individually!
Here are a few selected elevator pitches from last year’s cohort.
Even though elevator pitches do not contain slides, it’s a good idea to use slides to help organize and prepare your pitch. Here’s one approach:
- Create 4-5 slides, each corresponding to a single idea (perhaps one slide for each of the four pieces above).
- Each slide should contain a title that is a statement of an idea, and a single drawing or explanation.
- Practice, practice, practice. When practicing:
- Talk through the ideas that each slide represent as if you’re explaining them to a listener.
- Each time you practice, try to vary the way you explain each idea.
- Over time, your reliance on the slides will decrease. However, do not write a script and memorize it; that’s not engaging.
- Practice giving the pitch to members of your group. Let them interrupt you with questions. Can you adapt your pitch into a conversation if people ask you questions in the middle? If so, you have a strong pitch.
- Stand up, speak slowly, and enunciate!
You’ll present your elevator pitch in two stages:
- First, you’ll deliver a dry run of your elevator pitch to your TA and the rest of your project group. Think of the dry run as a “checkpoint” for your final elevator pitch – it doesn’t need to be polished at this stage. You will present your dry run at your Week 9/10 TA check-in; we’ll make an announcement on Ed once you can sign up for these.
- Then, you’ll individually record a video of you presenting your elevator pitch to YouTube and submit a link to it on Gradescope. There should be no slides in the video, just your face talking. (Let Suraj know if you have any concerns with this requirement.)
Note that every member of your group will have to record separate videos, even if you write your pitches together and they are largely similar.
Unlike the written proposal and schedule, your elevator pitch will be graded by your TA. Each of the six bullet points below are worth 1 point.
- Style: Between 1 and 3 minutes long and clear, and is not too rushed or too basic.
- Style: Not overly scripted, though reading off of notes a little is fine.
- Setup: Introduces the problem area to a general audience.
- Setup: Explains why the area is interesting to a general audience.
- Approach: Describes the specific problem being approached.
- Approach: Describes how the problem will be approached and how to measure success.
To assign a grade to your elevator pitch, your TA will use the following scheme:
|A||5 - 6|
|B||3 - 4|
|C||1 - 2|
All deliverables will be submitted to Gradescope.