The problem CovIntel solves

• The problem of finding the reach of vaccine campaign right now
  • by using APIs for integration to collect the data from reliable sources every 24 hrs.
  • by using maps to visualize the extent of vaccination carried out in an easier manner.
• Is the vaccine as effective as it was in clinical trials?
  • by predicting new cases over a duration, for a given geography and given the vaccine.
• Is vaccination finally leading to herd immunity?
  • by considering various factors like co-morbidities, vaccine strains and average age group in a region.
• Finding the optimal time between two shots for a vaccine
  • by monitoring the time period between multiple shots of vaccine and suggest an optimal time period for each vaccine.

Challenges I ran into

• Running the machine learning model over the backend and integrating with our app was an issue. We eventually generated CSVs from the ML models and kept updating the values and used the CSV files.
• We initially ran into a problem of what models to use to predict the cases in a region. One of the larger dilemmas we had were to go with pandemic models like SEIR (which did not show much correlation with CoVID in general) or probabilistic models like the ones we used. We cleared our dilemma seeing the accuracy of predictions our models were showing over a timeframe of 10-15 days.