CS-UDT flood digital twin

A crowd-sourced urban digital twin for hurricane flooding in Jacksonville. The storm replay is simulated, modeled on Hurricane Irma (Sep 10–12, 2017), and all community content is synthetic. Official station locations are real but approximate, and the citizen client reads your own phone’s sensors and location only if you allow it.

• Planner dashboard

  The twin’s map view. Press Play to run the simulated replay and watch reports converge into confirmed events. Best on a large screen.

• Citizen client

  The zero-install client residents would use: report, take quests, verify. Best on a phone.

What the demo loads and asks for

• The dashboard loads MapLibre GL JS from unpkg.com and map tiles from OpenStreetMap (© OpenStreetMap contributors).
• The citizen client loads qrcode.js from cdnjs.cloudflare.com.
• The citizen client can ask for your camera (to attach a photo), your location and your motion sensors, each only when you tap the control that uses it.
• Your demo reports stay in this browser (localStorage) until you clear the site’s data, with any location you attach. A photo is shrunk in the browser before it is stored, which also removes its camera data (such as where it was taken). If storage still fills up, photos stay only until you close the page. The demo sends nothing to this site.
• If you show a report’s relay QR code, anyone who scans it receives the report’s text and location (not the photo).

Background: Civic AI: crowd-sourced urban digital twins · CS-UDT build.