A number of volunteers (product, software, management backgrounds) are trying to build a model to predict how Oxygen demand will shape. When supplies improve, they wish to be able to help agencies distribute to the correct places, logistics issue as you might imagine. There is a model they are trying to build but too many unknown variables/uncertainties.

We are trying to build a model to predict demands on covid relief resources in coming days. Whenever supply side opens up (it will), we need a good way to distribute the supply to fulfil the demand. Here are the resources we should predict demand for: Oxygen cylinders, Oxygen refillers.

Supply-side bot

To get direct updates from suppliers on the availability of resources like medicines, oxygen, hospital beds etc.

Action plan (evolving)

  • These predictions can be made at various zoom levels: Country , States , Districts , Ward , Hospitals / care centers.

  • Build what if scenarios based on next few weeks and make recommendations

    • What if some districts, state manage to curb covid outbreak

    • What if one hospital doesn't have medical staff any more

    • What if covid spread increases

    • What if +ve diagnosis to hospitalisation ratio changes

    • What if lockdown happens partially?

    • What if virus transmission rate changes?

    • What if quarantine rate changes?

    • Impact of vaccination (reducing or increasing) spread.

Immediate goals

  • Estimate total Oxygen supply needed for Bangalore based on current data (no forecasting)

  • Ward / hospital wide distribution of Oxygen when supply is available


Excel model made so far

Demand data sources

Supply data sources:

What we need:

  1. Data Science folks: Relevant forecasting/demand prediction experience could be immensely valuable. We can share whatever data points we have. But most people in this sub probably already know what this entails.


Viksit Gaur Padmini Ray Murray Sumit Datta - Git URL Shreenath Regunathan

Last updated