enter image description hereI am quite new to QGIS and I am trying to create a species distribution map using public reporting data. I can very easily generate a heatmap by first converting the pentad data to vector centroid points, and then changing the symbology from "Single Symbol" to "Heatmap", and selecting my attribute of interest to weight them.

The problem I am having now is that I need each layer to be a vector, but the built-in heatmap output is a raster. Is there a way to do this conversion directly on the heatmap output?

When I try reclassifying using the processing toolbox > raster analysis > reclassify by table, my heatmap layer was not selectable as an option for the input Raster layer.

  • 3
    Changing the symbology does not produce a new map layer with which to use processing tools. There is also the tool Heatmap (Kernel Density Estimation)" in the Processing toolbox which will produce a raster on which you can perform further processing.
    – Matt
    May 21, 2023 at 8:58
  • Okay, thanks Matt. So even though I've changed the symbology, the layer has not undergone any conversion and is actually still a vector? With this being said, could you perhaps provide some guidance (or perhaps link to a previous thread) on how to use the tool Heatmap (Kernel Density Estimation) to generate discrete categories, as per my attached image? May 21, 2023 at 9:30
  • 1
    Can you elaborate a bit? Are the black dots your only factual points? If so, there are too few to take any notice of a heat map by probably 2 orders of magnitude. If you only have widely spaced and rare point data then I think you are better to use symbology that shows the magnitude at the exact points rather than attempting to model the continuous surface they might represent. May 21, 2023 at 9:41
  • Yes, the underlying layer is still a vector when changing the symbology. Symbology does not change the data model, it is just a particular visualization. So to create a "real" raster, proceed as recommended by @Matt - by the way: you should add this as an answer.
    – Babel
    May 21, 2023 at 9:41
  • Apologies for not clarifying @LeighBettenay. The data of interest is the purple banding - the black dots are irrelevant for now. This is how the data looks when the heatmap symbology is applied, and I am interested in getting it to a place where each band can be exported as a single vector layer. The best route is probably to explore the suggestion by Matt and Babel to first convert to a "real" raster and then proceed from there. May 21, 2023 at 9:55

1 Answer 1


Changing the symbology does not alter the underlying data. So, although the styling makes the layer appear as a raster, the source is still your point vector data.

There is the tool Heatmap (Kernel Density Estimation) in the Processing toolbox which will produce a true raster on which you can perform further processing.

Depending on your intended use, the output can be vectorised immediately, using the Polygonize (raster to Vector) tool. Though, this may produce polygons of a range too narrow to be useful (the polygons are essentially 1-pixel-wide contours).


enter image description here

If your intended use of the layer is for display purposes only, these narrow polygons can be styled in such a way to resemble the Heatmap symbology in your question by using a graduated renderer and making the stroke colour match the fill colour of each class:

enter image description here

As per the great suggestion by AWGIS, the Contour Polygons tool can be used to very easily create polygons of a given range from the Heatmap raster. Check the values of the raster and choose a suitable number for the interval parameter.

enter image description here

Alternatively, for more control over the ranges, you could apply an intermediate step to reclassify the raster into broader classes. See this question and answers for some methods of achieving that. The output of which can be vectorised by running Polygonize (raster to Vector) on it.

  • 1
    You could also use the "Contours" tool which allows you to set your start point and also your contour interval
    – AWGIS
    May 22, 2023 at 13:09

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.