I have a data set which contains NDVI (Normalized Difference Vegetation Index) values of four directions (NE, NW, SE, SW). Now, I would like to have single value which considers all 4 directions. Is it fine to take an average or is there any other metric that can be considered? Any suggestions on which is the appropriate method?

enter image description here

More information on the data source:

This data is from a competition from Driven data

NDVI is Satellite vegetation - Normalized difference vegetation index (NDVI) - NOAA's CDR Normalized Difference Vegetation Index (0.5x0.5 degree scale) measurements

  • 1
    $\begingroup$ I'm not very familiar with NDVI and after some googling I have no idea what these directions mean in NDVI sense. What is your knowledge on the origin of this data? $\endgroup$
    – Communisty
    Jul 17 '17 at 9:25
  • $\begingroup$ I have added more information on the data source $\endgroup$
    – Learner
    Jul 17 '17 at 9:43

After given some information about the data source I have a simple answer. Since the underscores SW etc. represent the four pixels around the city center and all the pixels cover the same physical area, taking a mean of the values will represent the average NDVI of the total area.

It is all about what you will do with the NDVI data. E.g. if something interesting occurs when somewhere near the center vegetation amounts are low (high), then the minimum (maximum) of these might provide you with more useful predictor.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.