12

I've spending last two weeks trying to discover how to do a Manhattan Voronoi on QGIS. For the difference between Manhattan Voronoi and "normal" Euclidean Voronoi, please check this link

Can someone point me to the right direction? In the link bellow you can find a sample of the information I've got (points and one polygon)

https://drive.google.com/file/d/0B4WB0ixNBUYkREtlcW0weDJUYU0/edit?usp=sharing

I would like to do in on QGIS, but if it's not possible, I'm open to other solutions.

6
  • 1
    Are you trying to create actual drive time sheds based on a road network, or just cool looking graphics based on pixel grids? Pixel grids would be as simple as locating your points within cells, then doing a raster calculation or query to find the MIN(ABS(X1-X2)+ABS(Y1-Y2))
    – user15741
    May 21, 2014 at 15:30
  • The points are zip codes centroid points. What I'm looking is to get the areas affected by each zip code, so answering your question, I guess I'm trying to create cool looking graphics. I'll going to follow your advice and I'll google how to do raster calculation in QGis. Thanks.
    – f616
    May 25, 2014 at 13:40
  • Zip codes are really irregular so Voronoi or Manhattan Voronoi will not represent the contents of the zip code very well. You could get a shapefile of all zip codes and if you wanted to simplify it while keeping it somewhat accurate you could convert it to a raster. You could then convert the results back to vector giving a simplified, roughly squarish zip code map.
    – user15741
    May 28, 2014 at 19:55
  • Indeed you're right, but portuguese postal office doesn't provide shapes of zip codes (long 4+3 format). However, for this propose it's not very relevant if the voronoi shapes matches, or not, the real zip code areas. The important is that the voronoi should be manhattan style to be more compliant with with street geometry. This voronoi will be used for pricing some telecom services. Once again, thank you
    – f616
    May 31, 2014 at 21:00
  • Could the Grass function v.net.centrality be relevant in this project? grasswiki.osgeo.org/wiki/Vector_network_analysis
    – ragnvald
    Mar 17, 2015 at 6:13

1 Answer 1

2

Create a dense fishnet (grid, mesh, etc) at a very high resolution. For each point in the fishnet, calculate distance to nearest of your points, but use Manhattan distance as the distance calculator. It will be an approximation, but presumably you only need parcel level detail.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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