Skip to content

Instantly share code, notes, and snippets.

@18dew
Last active April 4, 2020 09:40
Show Gist options
  • Save 18dew/b6d45de8f30dc70e3d8d69e522a37b40 to your computer and use it in GitHub Desktop.
Save 18dew/b6d45de8f30dc70e3d8d69e522a37b40 to your computer and use it in GitHub Desktop.
Covid Score - A Problem definition for tracking covid spread.

Covid Score

Problem Statement

In vast country like india its virtually impossible to test entire population hence we need to develop a solution which derives a risk score for public based on their exposure to infected people over a specificed period of duration. If the exposure is above a threshold limit specified by goverment then person will have to take test.

Purpose of the problem definition is to come up with a scoring mechasim based on intersection and exposure with infected person path.

Ideal scoring mechanism would quantify the level of exposure a person have had with infected person. ( Note the exposure has no upper bound )

The purpose of quantification of exposure to help optimize utilization of resources.( i.e. use of medical test kit, medical facilities, reduce panic among people at psycological level )

Axioms

All people in city keep a track of their locational data on their cell phone via an app and his locational data is broadcasted only if person is founded infected with covid to all the people of the city. ( All while respecting privacy of users )

Inputs

Following information is collected from every cell phone

  • Time ( at a Interval of 5 Seconds )
  • Location ( Latitude , Longitude and elevation is a bonus )
  • Accuracy of location ( on a scale of 1 to 15)

Broadcast information of infected person path is received based on diagnosis from testing.

( Information is not perpetual stored but for last 15 to 30 days )

Processing

There are two data sets that would be in use ( user path and list of infected person path over 14 days of time )

User path is compared with infected person path to identify following points

  • Number of intersection between users path and infected persons path. ( Note Infected person hold precedence i.e. - infected persn should have passed thne user should have intersected his path )
  • The duration of exposure ( i.e. - How much time user spend in and arround infected persons path )
  • Decay of exposure ( i.e. - If user is at same place with infected person and at same time then there is no decay of exposure but if user crosses infected persons path 5 days later then there is relativly less exposure )
  • Accuracy of location (i.e. - if Location is exact then the probabiltiy comes into play )
  • Elevation ( This affects only dense urban regions ) ( i.e. If person is on 5th floor and infected person is on 30th probability of infection changes gradually ) [ Factoring Elevation is optional ]

[ Note patients in quarantine are treated as infected lesser weight with a lesser weight ]

Final Output

There is only one final output

  • A Score denoting amount of exposure a user have had with infected people.

Use Case

Initailly goverment (i.e. - NCDC) would monitor data all exposed and then determine threshold for public based on onground observation via medical teams.

Post which NCDC would give four rating parameters

  • Green ( Clean ) - Person is free of covid despite minimal exposure
  • Yellow ( Self Quarantine ) - Person has minor exposure and needs to self quantine for 14 days and see if sysmtoms develop or not
  • Brown ( Conduct test ) - Person has major exposure and needs to urgent testing.
  • Red ( Hospitalization ) - Person does not need testing and must be treated for Covid as exposure is massive.

Implementation

In an ideal scenairo it should be an app which can give user information about their exposure. But in extreme case goverment should be able to identify infected people who need quarntine , testing and immediate hospitalization to prevent major calamity.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment