[Discuss] New database/web service: share S-R trends and estimated parameter values

[lisphilar]

Summary of new feature

We need to create a database/web service to share S-R trends and estimated paramer values of phases in all countries.

To speed up analysis of relationship of policy measures and estimated parameter values, we need to estimate the parameter values of phases in all countries. However, parameter estimation of models takes long runtime and records are updated every day becuase COVID-19 outbreak is ongoing.

Solution

1. Run PolicyMeasures.trend() and PolicyMeasures.estimate(cs.SIRF) with continuous Integration (CI) tools
2. Create RESTful API to open the estimated values

Note: CI tools limit on runtime of CPUs for free accounts.

Alternative

1. Create RESTful API to get/post estimated values
2. Volunteers run PolicyMeasures.trend() and PolicyMeasures.estimate(cs.SIRF) for some countires
3. Volunteers post the estimated values to database with the API
4. Users get estimated values from the database

[lisphilar]

Example of database/web service:

• GitHub
• Heroku
• Datapane
• streamlit
• Django

Example of tools to create RESTful API:

• apywrapper

[lisphilar]

Alternative: Speed-up parameter estimation with optimization of ODESimulator, Estimator and MPEstiamator class of CovsirPhy.

[lisphilar]

To use continuous integration (CI) tools in the new web service, CovsirPhy was registered in Open Collective (community to find sponsors of open source projects) and waiting for approval by the host. https://opencollective.com/covsirphy

[lisphilar]

This will be disucussed in forum #435.