Connection Atlas

Connection Atlas began with a personal problem: I could remember the people I knew but sometimes lose track of how they knew one another, or forget small details worth keeping. A contact list stores individuals; I wanted a map of the relationships between them. I independently designed, built, and deployed the app between June and August 2026.

The product represents people, notes, and multiple layers of relationships, with local search and path exploration. Breadth-first search (BFS) finds the shortest chain of recorded connections between people. The algorithm gives the map a practical purpose beyond visualisation: it can help explain how two people are connected through the relationships a user has entered.

Engineering focus: graph modelling and BFS applied to an everyday memory problem.

Relationship notes are personal, so I chose a local-first architecture without a cloud database. Records stay in the current browser/device. Client-side JSON import and export make backups portable and explicit, but local storage has a trade-off: clearing browser data can remove the map, so users need to keep their own backups. An exported file should be treated as private, too.

The result is a deployed tool with public source code, not a service that needs access to my visitors’ real social networks. It brings together a personally motivated product idea, a graph-based interaction, and deliberate choices about data ownership in a compact, independently delivered application.