Tree Map is an application for urban tree mapping. It helps local communities and enthusiasts collect and publish data about trees and plants around them. The application has a database of trees, their coordinates, current state, photos, and changes of that data over time. The data can be used for various purposes, such as urban planning, environmental studies, recreational, and more.
The application is heavily inspired by the New York City Street Tree Map (870k trees), which is a great example of how tree mapping can be done, and Treezilla (1M trees), which is a similar project in the UK. There might be other similar projects, but none that you can run easily and use in your own city, as far as I know.
The application is fully working an used by the Yerevan Tree Map project, although some features are still being developed, see Feature Map. The application can be quickly deployed on any server using Docker Compose.
The backend is a REST API written in Rust, using the Actix web framework. The database is SQLite, used via async-sqlite. Rust is chosen for the backend to make it fast and safe, being able to handle a lot of requests in tiny environments.
The frontend is written in TypeScript, using the Svelte framework. The map is provided by MapLibre GL. The frontend is a Progressive Web App, which means it will be installable on a mobile device and used offline one day.
The app works with any S3 compatible storage. For computational heavy tasks, such as processing 360 street panoramas, we use AWS Batch.
The project uses the monorepo approach to keep all the code in one place, for simplicity.
The services (backend, frontend, proxy) are located in the services/ directory.
Architectural decisions are logged in the ADR folder.
You can run the app locally using Docker Compose, like this:
docker compose up
With this command you'll be running the application, the frontend is available at localhost:5173. The SQLite database will be stored in a Docker volume, and should persist container restarts. The database will be created automatically on the first run. To run a public service, this would need a CDN.
There is a fully working public installation at Yerevan Tree Map.
If you are an UI/UX designer, we would love some feedback on how to improve the user experience. The focus should be on people doing the mapping work in the fields: adding new trees, measuring existing ones, and so on.
If you are into machine learning and computer vision, we would love some help on automating the process of initial mapping. Manually adding trees is the most time consuming thing so far. We would like to use video records of streets from a car mounted camera, to automatically detect trees and add them to the OSM map. We have a ticket for this, please join.
- New York City Tree Map
- London's Public Realm Trees
- Melbourne's Urban Forest Visual
- Berlin's Tree Map
- Paris' Tree Map
Know more? Tell us.