Having been closely involved in the existing process, I already had a solid understanding of the workflow and challenges. This allowed me to move directly to mapping the process into a set of user stories defining the system's required functionality. After creating a prototype, I presented it to the board and organisers, and we refined it through a few rounds of feedback to simplify key interactions.

The user stories formed the foundation for both the data model and API design. The project was initially prototyped using FastAPI and Angular, but I transitioned to SvelteKit to improve development efficiency. The final implementation was built with TypeScript and SvelteKit, using a PostgreSQL database via Prisma ORM. The user interface is based on IBM's Carbon Design System, adapted slightly for the public-facing components.

The application was deployed with a GitHub Actions, Docker and self-hosted Debian based continuous integration and delivery (CI/CD) setup, ensuring minimal costs, reliable updates, maintenance and full control of all data. It was used from spring 2023 until it's retirement due to big structural and operation changes within the organisation in late 2025.

The system has significantly reduced administrative workload, cutting the time needed to set up and staff a watch from about an hour to just a minute. It also provides a clear overview of upcoming watches and open shifts, improving staffing reliability and communication with parents.

The following year GDPR consent for the Natteravn was omitted from the general consent form issued by the school to all parents at the start of the year. This made it difficult to use the system effectively as it relied on contact information being available. The challenges around GDPR forced a complete rethink and radical restructure of the way the organisation is run, making the system as it was obsolete. It was however an interesting project to work on and a good example of how even a small self-initiated project can cut down on administrative workload.