Visible Infrared Imaging Radiometer Suite (VIIRS) Series

VIIRS is a fundamentally better series than DMSP for measuring economic activity. It does not suffer from observable Top and Bottom coding and reports absolute radiance rather than relfectance, (for more details in data difference please see http://rpubs.com/parthakhare/216662) The series was formally released in 2014 and collects stable lights data at a monthly frequency. The last reported data was collected in August, 2016.Before releasing the formal VIIRS series ranging from 2014-2016, a Beta series was released for sample months in 2012 and 2013. In our present we want to proxy VIIRS for economic activity for India. The raw data is cleaned, resampled at 1 square km and extracted by Indian State Boundaries.

Plot Analysis: Understanding VIIRS Distribution for India

Monthly Night Lights (VIIRS) Variance| 2014:2016

The distribution shows the inter month variance, as well as providing a tool for understanding state based localised impacts where States do not confirm to the conventional behaviour. In November exhbits highest values whereas January shows the least.


Level Plots: Average Anuual Night Lights (VIIRS) ’14-’15

The average annual numbers for per capita numbers indicate growth across time. Telangana, Harayana and Punjab are trending in opposite to growth direction. Apriori small state seems to have something with it, for the same we are checking monthly plot

## Loading required package: grid

## NULL


Scatter Plots: Night Lights (VIIRS) & Gross State Domestic Product

Comparing Gross State Domestic Product with sum of Night Lights (VIIRS) values It can be seen that most of the states are confirming for 2014, 2015. The blue line is the regression and grey region represents 95% confidence region
The blue line represnts a fitted regression and grey region represents 95% confidence region


APPENDIX

——————————————————————————————————————-

Using additional years data for Long term growth analysis

In addition to 2014-2015 we would want to utlise 2012, 2013 and 2016 values, in dearth of a reliable long term series from DMSP. In this, we have tried to study trends adding data from the said years. The objective is to try and identify how more years of data can leveraged.

VIIRS Data Vinatge

VIIRS Monthly Data is present by: 2014,2015 (All 12 months), 2016 (January - August), 2012 (April, October), 2013 (January)

Monthly Level Charts: January [2013,2014,2015,2016]

Includes 2013 onwards data, January is seen to record usually the lowest values throughtout the year. 

## NULL

It is clear that the monhtly variance of this data cannot simply allow for interchanging months. For incorporating this data, two approaches have been considered, a.) Averaging each year’s values b.) Checking the trend only for given months for which 2012,2013,2016 reports values

Monthly Level Charts: April Charts [2012,2014,2015,2016]

## NULL