I have census block data that I have clipped using my 1/4 mile buffer. For each of my census polygons it has kept the original population, is there a way to get the population recalculated just within the buffer? enter image description here

  • What GIS software are you using? What precisely have you tried? – PolyGeo Oct 7 '16 at 16:01
  • 1
    I have seen this come up a few times and its ugly - think about if you cut a census block to 50% of its area - does that mean you take 50% of the population? Are you comfortable generalizing your analysis to say that population within each block is always evenly distributed? And - what about larger census blocks that have their population outside the area that was clipped? – DPSSpatial Oct 7 '16 at 16:04
  • The simplest way top avoid uniform distribution is adding address points to the analysis. Compute heads per address using census etc – FelixIP Oct 9 '16 at 3:42

As @DPSSpatial has alluded there is not a simple clean cut solution here. You are bumping up against a couple prominent issues, namely the Modifiable Areal Unit Problem and Spatial Heterogeneity.

Long story short--no matter what you do you'll be introducing new errors into your results. It's unavoidable, and inevitably involves some sort of interpolation.

One of the most simple is binary dasymetric (this is better, in my opinion, than say, areal weighting) areal interpolation.

Here's a couple of examples: http://www.colorado.edu/geography/leyk/data/Hamid/dasy%20lit/eicherBrewerDasy.pdf https://anitagraser.com/2012/11/18/improving-population-density-maps-using-dasymetric-mapping/

It was also recommended to me that I take the total population of the block vs the percentage of the block within the buffer and apply the math to the total.

  • Your remark at the end of my answer is an example of areal weighting. While that would improve your original method, it's still more error prone that the binary dasymetric interpolation method. – rumski20 Oct 7 '16 at 18:36

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.