The project is called GeoPrint and it was carried out with Supsi, Virginia Tech, EFREI students. The idea of this project is by CoStar group.
This website will allow the user to select a certain area on the google map, this selection will be processed by the ML code. Information about the building, such as its size, coordinates, and footprint, will be shown.
The possible targets could be:
- Real Estate: Retrieve property dimensions to provide accurate listings for clients, visualize footprint on map to showcase listings to buyers.
- Urban Planner: Generate footprint maps to plan infrastructure projects, compare current/past building footprints to analyze trends over time.
- CoStar Data Analysts
Users can zoom in and draw a rectangle around an area of interest with just two clicks. The system responds with pop-ups, and the frontend displays the detected building outlines and the area (in m²) for each building. All calculations are performed by the ML model, which also provides confidence scores.
Installation of some tools is required before starting.
in the website/ folder allows you to download all necessary dependencies
npm installhere the entire list of dependencies that were used:
- vitejs
- react
- react-dom
- react-router-dom
- react-icons
- tailwindcss
- typescript
After that, it is possible to run it using this command
npm run devFurthermore, it was put in the VT Cluster. Therefore, GeoPrint can also be used through its links:
- the main link https://propertyfootprintgeneration.endeavour.cs.vt.edu/
- the shorter link https://bit.ly/GeoPrintApp
The project cannot be used properly locally, because both parts, frontend and backend must be started at the same time using the docker. But you can start it using either of the two link given above.
