CovidCast is like a weather forecast; but for coronavirus outlook in your area. Due to the rapid spread of COVID-19, social interactions are to be kept to a minimum. Even doctor-patient interactions have been impacted. This poses a significant challenge for patients as they might not know if their condition requires in-person care. As a result, patients might not receive the care they need, resulting in a worsening of their condition. At present, leaving one's home poses a certain risk, but this risk is not so readily quantified. Our product tries to solve both these problems, while also being accessible to those with limited or no internet access.

What it does

Our project focuses on three core ideas: First, our “Covid Check” feature simulates a typical doctor patient interaction in order to help patients assess their condition and advise them - if needed - to seek medical attention. We hope this will improve patient compliance as well as keep patients safe!

Secondly, our “Covid Risk” feature allows the user to assess their COVID-19 exposure risk before they even reach their destination! Our risk analysis algorithm takes into account the occupancy at their destination (popular times), air quality index, and the number of active cases in the region. These parameters were chosen as current research suggests they play a significant role with regards to the spread and severity of the disease. Risk levels are categorized as either very low, low, medium, high, or very high. We hope to provide users with an accurate reflection of the risk they take during this global pandemic.

Thirdly, as we continued our research, we found that many people around the world, especially in developing countries, do not have access to reliable internet. We decided to make an alternative device that could collect the same information from the internet and then relay it to more rural areas over radio waves. By using a Raspberry Pi Zero W and a radio module, we could create this device for under $25 USD. It was imperative that our project be as inclusive as possible. Along with the Raspberry Pi device, our product is compatible with google homes and is ready for use.

How we built it

  • Created a COVID-19 checklist based on current standards
  • Developed an algorithm for the risk analysis using COVID-19 case data, air quality index, and destination occupancy (popular times)
  • Used JavaScript to code the inputs for risk analysis and generate a risk analysis.
  • Used DialogFlow to integrate our Covid Check and Covid Risk to Google Assistant
  • Integrated Raspberry Pi technology with Google Assistant features.

Challenges we ran into

Like any project, we ran into some challenges. The first challenge was that all of us had either little to no experience using JavaScript. We also discovered that the Google Maps API does not have a method to retrieve the popular times of a location. To help, we talked to mentors and did our research to come up with a solution.

Accomplishments that we're proud of

We are really proud of our project - all 3 components - and how we managed to do this within the 36 hours of the hackathon.

What we learned

We learned so much during this whole hackathon! It took us several hours to learn how to write the code and get the right APIs for the risk analysis. We all had very limited experience with JavaScript and working with DialogFlow, so it was really cool understanding and learning how they work. It has been really great to learn about research through published papers, coding, and technology that we have not used before!

What's next for CovidCast

We have many ideas on how to improve CovidCast. For our risk analysis algorithm, more parameters could be added (ex. socio-economic factors) and machine learning could be employed. Added language support would help us better serve those who do not speak English. In addition it is ready to add to a Google Home and we can work on developing the offline interface as well.

Share this project: