Project Proposal (15%): The proposal is where your group will (unsurprisingly) pitch its project idea. The goal of this stage is two-fold. First, it is a good initial, yet relatively low-stakes test of your group dynamics. As this is where you will introduce the problem, methods/solution, data, and so on, such a task requires consolidation and universal contribution to the overall vision of the project; in a word, teamwork. And second, the goal of the proposal is to get early-stage feedback from me on the feasibility of the project, both regarding your ability to complete it within the time frame of the course, as well as whether your project is in the scope of the class.
There are relatively few guidelines for the subject matter of your final project. My only requirement is that the methods at a minimum incorporate machine learning techniques covered in the course. For example, you could use a clustering technique to label (previously unlabeled) data, which you could then feed to a supervised classifer to predict future classes. Or alternatively, you could fit a variety of supervised classifers (e.g., random forest, penalized logistic regression, neural network, SVM, etc.) aimed at classifying the likelihood of some candidate winning political office based on prior election performance, and then compare quality and accuracy across the algorithms.
You will be working in a team of ~ 2-4 students, which you will organize on your own. Groups must be organized by the end of the first week of class (Friday). Note: If you are having trouble finding a group, reach out to me sooner rather than later to request help and I will assign you to a group; if I don't hear from anyone, I will assume all students are in groups. As time will go very quickly, you are strongly encouraged to keep in regular communication with your group members, whether in person, virtually (e.g., Slack, email, text, etc.), or both throughout the quarter. If you have any concerns working with someone in your group, please talk to me for accommodations.
Some particulars: ~ 2-6 pages, excluding references; it is recommended (not required) that you use figures to illustrate technical concepts. Final submissions should be a single PDF document rendered from your Markdown (or Jupyter Notebook) document. Use 11- or 12-point standard font and include a properly formatted reference list. You can select any reference style you wish (Chicago, APA, MLA, etc.); just be correct and consistent.
Note: team proposals are due at the end of Week 3 on Friday.