The main concern in the door-to-door vaccination process was the task to consume vaccine vials within the limited time . Since after opening a Covid-19 vaccine vial, all doses in it have to be administered within four hours. Else, they will go waste and have to be destroyed which will be a huge loss to the current situations of the country.
A vaccine vial contains a certain number of doses and to avoid wastage we need to use a maximum amount of it.
Another problem in the door-to-door vaccination process is the travel time required to reach one house to another.
The vaccination process can be made user friendly by sorting and grouping registered people on the basis of their location in a defined range i.e. cluster formation(K-Means)
And confirming their registrations only when we receive enough(doses in a vaccine vial) registration from a given area so that no or minimum wastage of vaccines occurs.
This will help us solve the hurdles of wastage and time taken in travelling as we will have to cover places which are near that are within the defined range .
There were a lot of algorithms we tried to implement theoretically and the best one we found was the k-means clustering method . Yet in this method also some hurdles are there like we need to define the number of clusters. We researched some youtube videos and then found a proper method to define this k value. This was achieved by checking the change in variation we get in each successive value of k, and when we get an elbow point (see fig). All this coding was done in JavaScript.
Even after getting k value we still have to properly plot all this data on the map. Finding all the map APIs and then going through it’s documentation was tiring work. Since we wanted to run the entire application properly, we needed to ensure every api, database and javascript programs work properly with the frontend. We will achieve this by properly dividing all the work between all team members.
Technologies used
Discussion