The problem Covid-Pred20 solves

To control the spread of COVID-19, screening large numbers of suspected cases for appropriate quarantine and treatment is a priority for the government.

Fast and accurate diagnostic methods and tracking travel history of suspected are urgently needed to combat the disease. Currently, there are only 62 out of 19,817 hospitals dedicated to COVID-19 treatment. Hence if a pandemic hits on a large scale we require a preliminary test using existing accessible technology instead of new technology.

Detecting the travel history of 14 days for every patient is not feasible, hence we need to automate it using structured data.

Challenges we ran into

Training, evaluating and infering our complex deep learning model with hugeeee dataset. Also, deciding how to collect data from the customer side(citizens and doctors).