(presentation)=
# Final Project: Lightning presentations

## Logistics

- ~4-5 minutes (max 5 minutes) per group
- Format will be Google Slides to make transitions between groups easier
- Sign up for a presentation slot [linked here](https://docs.google.com/spreadsheets/d/1e3F6CNbZeeMMwMAY-RBxDBaXy7dgVkr1Yz6-4M2poFY/edit?usp=sharing)
- [Presentation template](https://docs.google.com/presentation/d/1jLwe2jX7uT0yP_jYxUqZ-E-ZAfDqXikUlrY9rHaY2Xc/edit?usp=sharing)
- [Shared presentation slides](https://docs.google.com/presentation/d/1Gr82GfMzLDr3P3X6WDVVAhaQ_hfF9dxKxgUmxuOu_xs/edit?usp=sharing)

## Guidelines

- You should cover your ML question, dataset, and a baseline metric. Your presentation does *not* need to include fully tuned model performance results.

- **Motivation**: Introduce the problem to a general audience.
    - Think of yourself as pitching the significance of your prediction task. Why does this question matter, and what could someone do with a good prediction?
- **Dataset**: Introduce your dataset and how it connects to the ML question.
    - Describe what the instances and features represent, what the prediction target is, and any notable characteristics (size, class balance, sensitive attributes).
- **Evaluation metric and baseline**: Describe your evaluation metric and briefly discuss why you chose this metric for your evaluation. Then, provide a baseline result on the test set.    
- **Next steps**: Conclude with a brief forward-looking statement about your remaining modeling and data engineering plan.

## Peer feedback

During the presentations, you will give feedback to your classmates according to the following template:

Please provide one constructive point (though feel free to add more) for the two sections below. Your feedback will help the presenting group improve their final submission.

### What I enjoyed

Things you can consider:
- Clear motivation for the prediction task
- Interesting or well-chosen dataset
- Thoughtful consideration of who would use this model and who it could affect
- Clear description of features and target variable

### Thoughts on dataset and approach

Things you can consider:
- Are there any potential issues with the dataset (e.g., class imbalance, data leakage, missing values) that the group could address?
- Are there features that might be sensitive or could introduce bias?
- Are there additional preprocessing steps or alternative models the group could consider?

## Rubric

| Section | Points |
|------------------------------------|-------|
| Motivation and ML question | 1 |
| Dataset: features, target, and characteristics | 1 |
| Baseline metric and next steps | 1 |
| Peer feedback | 2 |
| Total | 5 |