Project Management & Sprint Planning

Agile Approach

From day one, Planwise embraced an agile, sprint-driven workflow to keep our team focused and adaptable. This approach enabled us to incrementally deliver value while maintaining flexibility to incorporate feedback and new insights.

Sprint Planning & Execution

Sprint Backlog & Goals

We kicked off with a sprint planning where the Product Manager organized a group meeting to set goals and distribute tasks. High-level objectives were translated into small achievable incremental goals, allowing us to move from an MVP to the final delivery.

Our process was managed on Trello, where team members could:

• Choose tasks during each sprint
• Mark tasks as completed
• Move on to new tasks as previous ones were completed

Tasks were distributed to play to each team member’s strengths and preferences across:

• Data engineering
• Model development
• Front-end development
• Back-end services

This approach enabled simultaneous progress toward our MVP and final product.

MVP Delivery

By the end of the first sprint, we launched a functional Streamlit interface where users could:

• Adjust preference sliders
• Receive a ranked list of Madrid’s top spots
• View recommendations powered by our initial autoencoder model

This early prototype validated our end-to-end pipeline and set the stage for rapid feature expansion in subsequent sprints.

Development Workflow

Version Control

We established a branch-based workflow with:

• Main branch reserved for stable releases
• Feature branches for all new work
• Pull requests with thorough code reviews
• Issue tracking for bugs and feature requests

Code Quality Practices

To maintain high standards of code quality:

• Comprehensive docstrings and type hints span the entire codebase
• Pre-commit hooks enforce Black formatting, organize imports with isort, and catch linting errors via flake8
• Unit tests cover data-pipeline logic, model training and inference, and API endpoints
• Coverage reporting ensures code is adequately tested

Deployment Pipeline

Our CI/CD pipeline automates:

• Linting and style checks
• Type safety verification
• Automated testing
• Docker builds
• Deployment to staging and production environments

This approach ensures that every change is thoroughly vetted before reaching production, giving the team confidence in each commit.