Visit the iCovid Website
Inspiration
We were inspired by the current pandemic and how simple, yet powerful analysis and ML tools could help us understand this pandemic better.
What it does
- Allows you to see the growth of the virus for a selected country.
- Get predictions for growth in the future
- Monte Carlo simulation with different parameters showing how variants can affect virus propagation
- Augmented reality world map with covid cases
How we built it
We used Vuejs for the frontend with libraries such as 'leafy' to display maps and 'vue-chartjs' to display charts. We built the backend using Node.js and Express.js. The AR was built from the ar.js library. Tensorflow was used to predict and forecast future COVID19 cases.
Challenges we ran into
It was hard to connect the backend and the frontend when deploying with Heroku. We also had trouble displaying a Doughnut which we abandoned at the end. There was a problem with tensorflow.js since a model trained with Keras was not compatible even with conversion using proper tools.
Accomplishments that we're proud of
- Deploying a web app on backend and frontend with Heroku
- Even though the AI predictions were not possible on the browser, the results were still satisfying when running on a local machine.
- Using AR to build a dynamic map with the latest COVID-19 cases around the world.
What we learned
- While what we did was very interesting, it opens up possibilities for further analysis and predictions.
What's next for MariHacks 21 Project
- Have the AI implemented on the browser.
- Make the simulation more visual by showing a heat map of the cases.
- Fix a few bugs related to the displaying of the data.
Built With
- express.js
- heroku
- node.js
- tensorflow
- vue
Log in or sign up for Devpost to join the conversation.