COVIDcast

Project Introduction

Due to the ongoing COVID-19 pandemic, leaving one's home poses a potential exposure risk, but this risk is not so readily quantified. We have developed a solution that provides users with a forecast of the COVID-19 outlook in their area. Much like how one quickly checks the weather forecast before leaving home, it is our hope that users will check COVIDcast before leaving for their destination. The primary aim of our application is to keep users safe and healthy. Accordingly, we present users with a list of alternative destinations that may pose a lower COVID-19 exposure risk based on real-time data.

Issue/Inspiration

Leaving your home poses a COVID-19 exposure risk, but this risk is not so easily quantified. We would like to develop a solution that allows users to easily assess their COVID-19 exposure risk in a manner that’s easy to understand and access similar to checking the weather forecast before leaving the house.

What it does

The application determines the COVID-19 risk assessment level associated with the destination. This risk level is based on the number of active cases in the region as well as an estimate of how busy the destination currently is (Google Popular times). It also helps the user find alternative nearby destinations that may pose a lower COVID-19 exposure risk.

How did your project evolve with the support of the COVID-19 hackathon fund by Google Cloud?

Originally, our project started off as a Google Voice Assistant application. We shifted our focus towards developing an Android application. The reason for this change in platform is that we believe smartphone users would be more engaged and it enables us to readily develop more features.

How you built it

The original Google Voice Assistant Application is powered by the language processing abilities of DialogFlow. User’s ask for the COVID-19 exposure risk for a specified location at a specified time. In turn, the application calls a webhook service (Firebase Cloud Function) which retrieves the COVID-19 exposure risk level. This risk level is based on the number of active cases in the county and the anticipated number of people at the destination (based on Google Popular Times). A second intent was developed that allows the user to perform a COVID-19 self-assessment. Depending on the responses from the user, the Google Voice Assistant application may recommend that the user get tested for COVID-19, seek immediate emergency medical attention, or continue to take precautions to avoid exposure to COVID-19.

Our Android application was built in Android Studios. The Places Autocomplete service is used to find the destination of interest. HTTPS callable Firebase Cloud functions return the number of active cases in the county as well as an estimate of how busy the destination currently is (based on Google Popular Times). We have also created an HTTPS callable Firebase Cloud function which finds alternative nearby locations with the same Google Place type that may pose a lower COVID-19 exposure risk. Color coded markers are placed on the map which indicate the risk level of the destination.

Challenges you ran into

Managing work on COVIDcast alongside our regular school duties and other responsibilities. Integration of cloud services, especially given that this was our first time using Google Cloud services First time programming in Javascript and self teaching the language

Accomplishments

We are very proud of the progress we have made on COVIDcast since the Hackathon. We are proud of developing a meaningful solution in the fight against COVID-19. This was also the first hackathon many of us had participated in!

What you learned

How to develop callable Firebase functions How to call Firebase functions with an Android device and send the data collected from Firebase functions to an Android device. How to perform Place searches and apply filters (ex. Place type, open now…) Creating Dialog Flow Intents and adding webhook fulfillments How to write basic code in javascript

What’s next for your project

We hope to expand the scope of our risk assessment algorithm. Additional parameters we are considering incorporating include the anticipated visit duration, vaccination rates, age, regional masking regulations, proportion of cases that are variants of concern… We also hope to develop a COVID-19 risk assessment algorithm for public transportation We aim to complete the Google Voice Assistant Application we originally worked on during the Hackathon. We intend to release the Android Application on Google Play in the coming months and eventually support more regions

What Google Cloud products did you use to build your project?

• DialogFlow: original Google Voice Assistant Application
• Firebase: callable functions
  • Similar locations: perform a nearby search (Places API) of similar locations (matching Place type) and return an array of max 20 locations
  • Get active cases: return the number of active cases in the county. Web scrapping of the NYT COVID-19 repository
  • Get Popular times: returns the popular times information for the destination. This also uses the Places API
• Maps: draw the map and place markers on the Android application
• Places API: perform nearby searches, Places Autocomplete, finding the Popular Times