1

I'm having a hard time making the heatmap plugin do precisely what I want (namely, to make a heatmap showing not only points but an attribute of those points), and now I'm thinking that I may be misunderstanding what "min/max of full extent" means.

To my mind, the minimum and maximum values QGIS pulls out of the attribute field and displays in the band rendering section should be the lowest and highest numbers that appear in that column. Is that actually correct?

(In that case, my max should be 6500 or so. Instead, it's 10.35.)

  • I'm not exactly clear on what you're asking. I read a couple of heatmap tutorials, and it seems this is about symbolizing the heatmap results, correct? Does step 12 at this tutorial answer your question? The heatmap plugin appears to be about concentration of points (density) and it sounds like you want value contours (interpolation). Your meaning of min/max full extent is one possible correct definition, but you might be off with the application (which field) in this case. – Chris W May 17 '14 at 19:13
1

The answer is a little bit technical... What actually happens is that the value from each point (usually 1, unless you're using a weight column) is "spread out" over the chosen heatmap radius. Imagine a spike of height 1 at each of your points which has been squashed down to make a bump shape. Each of these is then added to get the resultant raster.

To summarise:

  • the maximum value that each feature contributes to your raster will be less than the attribute's value, since this value is "spread out" over a circular area.

  • the smoothed value from each feature is added for every pixel in the raster, so the maximum value in the raster will also be affected by the density and number of points in your table. Thus the maximum value in your output raster will depend on both the radius/smoothing type applied, and the spatial distribution of these points.

  • pixels in your output raster which are further then the search radius from the nearest point will have a zero value.

Google "kernel density estimation" if you're after more details... that's what the mathematical process behind heatmap creation is called.

  • Thank you so much. That explains a lot! It sounds like I might be best served by mapping duplicate points. I'm trying to show the distribution of production capacity, so I'm going to try creating a duplicate point to represent each unit produced at each location. With the duplicates, I'd need a straight-up heatmap, which should be easy to make with the plugin. – Dani May 19 '14 at 14:58
  • @Dani I really don't think that approach is the most direct route to what you're trying to do, nor will it yield the result you're looking for - density and interpolation of values are two very different things. Do you really want 6500 points stacked on top of each other? I strongly suggest reading about interpolation and then taking a look at the IDW tool in QGIS. – Chris W May 19 '14 at 22:15
  • I came to the same conclusion as I made the duplicates; clearly, this can't be the best way to do anything, but I wasn't sure what other terms I should Google. THANK YOU! I'm going to go read up on interpolation right now. – Dani May 21 '14 at 14:40

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.