General Advice

[pjpaulpj]

Hi Parth,

I'm an economist interested in using the VIIRS data to proxy for economic conditions. I found your breakdown of DMSP v/s VIIRS quite informative.

I'd like your advice on the suitability of these data sources for policy analysis. One take-away from your post is that there's a high degree of month-to-month variability inherent in the data. I'm imagining that this would make the data less sensitive to policy changes-- say a new SEZ or factory in a region. Is this your read as well?

On a related note, would you have suggestions to get started on the sort of analysis that you perform here? Any tutorials or books that you found helpful?

[parthkhare]

Hi,
Glad to connect!

To start with, VIIRS is definitely better than DMSP but as one has to sift through the various strands of nightlights imagery (viz. monthly, yearly). I had observed varying results just calibrating monthly and yearly estimates itself.
In terms utility for policy, broad brush stroke validations can run well with the cloud cover cleaned yearly imageries however for small area statistics, the raw data mandates contextualised cleaning and application. Would be glad to discuss any problem statement you have in mind.
It's almost an oxymoron to say my learning was old school, in that I have scrambled pieces from gistackoverflow and NOAA mostly. But, offlate there have been succinct blogs on this, will share with you, as I know.