0

I am working with two data sets from different sources that represent the same phenomena and have a challenging point.

Both data sets represent demographic attributes of the population [ person id, age, gender, income, ... ] and records are matched to the zone number with no specific location. So I can only map them to the center of the zone.

My question is, what algorithm, method, technique is suitable to find a common grid system for both zones for comparison? They should represent a high correlation as they are for the same population.

I use PostGIS, QGIS, Python and Java

two zone systems overlaid

Matt
  • 16,843
  • 3
  • 21
  • 52
Mohamed Batran
  • 9
  • 1

1 Answers1

-1

Probably you can divide both the datasets to smaller grids and then run comparative analysis.

For making grids in QGIS refer to answer of this question.

For splitting them you can use fishnet in ArcGIS.

csk
  • 24,827
  • 3
  • 32
  • 70
Sandeep Kumar
  • 633
  • 4
  • 12
  • Thank you for the replay, I think the problem is that the whole records are only in each zone center. so, only the grid that intersect with the center of the zone will have 100 % of the zone records which which will lead to very low correlation – Mohamed Batran Aug 24 '17 at 09:58
  • but, one Idea I got from you comment is to reallocate or geo-smooth the point distribution in each zone based on some probability distribution, then continue the grid comparative analysis – Mohamed Batran Aug 24 '17 at 10:01