This repository houses a template for a shiny web application that uses a machine learning model to classify webcam images for flooding and collect training images and labels from users.
This template includes code for deploying the application to Google Cloud Run. This template is based off the COPE COMET "sunny-day" flooding shiny application, the NC12 Flood CamML.
The published web application is:
- built with R using {shiny}
- writing data to Google Sheets & images to Google Drive
- containerized with Docker
- hosted with Google Cloud Run
See the Flood CamML website for detailed instructions. The tutorial will explain how to use this code template, set up the necessary Google APIs/permissions, and deploy the app to Google Cloud Run.
An example (NC-12 Flood CamML) is available at nc12.floodcamml.org
Code for deploying the application to shinyapps.io is available in the FloodCamML_shinyapps repo.
CamML is an open source project for crowd labeling and ML prediction of real-time webcam imagery. See the full project description at floodcamml.org.