retrospective.qmd

---
title: "Project Retrospective"
---

## Overview

This retrospective documents our team's reflection on the TravelPy project for DSCI 524, using data from our GitHub Project board.

## Project Analytics

### Team Workload

| Team Member | Issues Completed | Primary Focus |
|-------------|------------------|---------------|
| hoomanesteki | 30 | CI/CD, Testing, Documentation |
| hugokwok0119 | 10 | `estimate_trip_cost`, Bug Fixes |
| derrickmichaelj | 10 | `format_destination`, Peer Reviews |
| RNanfuka | 10 | `get_packing_list`, Code Review |

**Total Completed:** 63 issues

The workload distribution appears skewed, but hoomanesteki handled many smaller CI/CD configuration tasks, workflow debugging, and documentation updates that were quick to complete. The other team members focused on more complex core logic implementation.

### Milestone Progress

| Milestone | Issues | Focus |
|-----------|--------|-------|
| M1 | 13 | Function specifications, initial setup |
| M2 | 4 | Core implementation |
| M3 | 17 | CI/CD, testing, documentation |
| M4 | 29 | Polish, peer reviews, final fixes |

M3 had more issues than M2 due to CI/CD debugging GitHub Actions workflows required multiple iterations to get right. M4 had the most issues as we addressed peer review feedback and finalized documentation.

## DAKI Analysis

### Drop
- Manual version bumping (switched to `hatch-vcs` with git tags)
- Large PRs that were hard to review

### Add
- Pre-commit hooks to catch linting issues before pushing
- Issue templates for all task types

### Keep
- Matrix testing across OS and Python versions
- Branch protection with required reviews
- Codecov integration for coverage tracking

### Improve
- Better time estimates for CI/CD setup
- More parametrized tests to reduce duplication

## Tools & Infrastructure

| Tool | Purpose |
|------|---------|
| Hatch | Package management, environment handling |
| pytest + pytest-cov | Testing and coverage |
| Ruff | Fast linting |
| Quartodoc | API documentation generation |
| GitHub Actions | CI/CD pipelines |
| Codecov | Coverage tracking |

### CI/CD Workflows

| Workflow | Trigger | Purpose |
|----------|---------|---------|
| `build.yml` | Push, PR | Run tests on multiple OS/Python versions |
| `deploy.yml` | Tags | Publish to TestPyPI |
| `docs.yml` | Push | Build and deploy documentation |

## Scaling Recommendations

For future projects like capstone:

1. **Automate early**: Set up CI/CD in M1, not M3
2. **Use pre-commit hooks**: Enforce standards before commits reach CI
3. **Keep PRs small**: Easier to review and debug
4. **Document as you go**: Writing docstrings during development saves time

## Key Takeaways

- CI/CD setup took more time than anticipated due to debugging workflow failures
- Work distribution by issue count doesn't reflect complexity explain context in retrospectives
- All 63 issues completed with 100% completion rate across all milestones
- No significant bottlenecks in review process