Support with District-Wise Vaccine Delivery - Throughput data per day

[aashir...@gmail.com]

Dear All,

Using the CoWIN API I managed to calculate the district-wise throughput data per day for vaccine delivery in India.

Throughput per day per district = (SUM(Number of First Doses + Number of Second Dose)/Total number of Sessions)*Total number of sites 

The only way I get this data to be usable is to append the throughput numbers with the weighted average of the population of the district. This will provide an accurate representation of the vaccine delivery performance per district (although this will still be at an aggregate level).

Can anyone point me to accurate population data per district as per the sheet enclosed as the attachment. It will be of immense help.

Sincerely,

Aashir.

[Bhanu K]

hi Aashir,

This is an interesting work, thanks for sharing... From what I gather, population projection data at districts is not available publicly. I discovered two datasets for projecting populations at state-level recently: 

1) UIDAI (PDF)

  - estimates as of 31 dec, 2020

2) NHM (big PDF)

  - estimates as of 1 mar, 2021

  - more granular estimates by category (age, gender)

I'm not an expert in this area, interested to learn. A friend suggested estimating population from raster images shared by the likes of WorldPop (another source).

Step 1 - estimate population

I've used QGIS (raster zonal stats) to estimate population. Tutorial for step-by-step guide. The files I used are from WorldPop, India districts shapefile and use sum instead of average metric in zonal stats.

After this step, we'll have a population estimate at district level.

Step 2 - compare against baseline

2.1 Aggregate the population numbers at state-level.

2.2 Use state estimates from UIDAI or NHM as a baseline.

2.3 Determine the relative change rate (raster estimated population divided by state estimate from UIDAI or NHM).

2.4 Use this rate and divide the district estimate (at the end of step 1) by it.

I've done this for states and shared the results here. For most states (with multiple districts) the approach seems to work well but fails for smaller regions.

If anyone else has done this earlier do share feedback on improving this approach or suggest alternative ways.

Thanks,

Bhanu

--
Datameet is a community of Data Science enthusiasts in India. Know more about us by visiting http://datameet.org
---
You received this message because you are subscribed to the Google Groups "datameet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to datameet+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/datameet/a677a25d-d560-44b3-8739-548b1feb4d82n%40googlegroups.com.