A heat map turns a cloud of points into a picture you can read at a glance: where things cluster, where they thin out, and where the hotspots are. Crime incidents, disease cases, shop locations, traffic accidents, species sightings, any point dataset becomes clearer as a heat map than as a scatter of dots. QGIS has a built-in heat map renderer that takes about two minutes to set up, and this guide walks through it step by step, with the settings that actually matter and the mistakes that make heat maps lie.

What a Heat Map Really Shows

A heat map, technically a kernel density surface, estimates how dense your points are at every location on the map. For each point, it spreads a smooth hill of influence over the surrounding area, and where the hills overlap, the values add up. The result is a continuous surface that is high where points cluster and low where they are sparse. The two settings that control it are the radius, how far each point's influence spreads, and the weight, whether some points count more than others. A large radius makes a smooth, general picture; a small radius shows local detail but looks noisy. The colour ramp does the storytelling: a good ramp goes from transparent through yellow to red, so the eye reads the hotspots instantly. The most important thing to understand is that a heat map shows density, not counts, and it does not show causation. A hotspot of reported crimes might mean more crime, or it might mean more reporting.

Setting Up the Heat Map Renderer in QGIS

Load your point layer into QGIS, then open Layer Properties and go to the Symbology tab. Change the renderer from Single Symbol to Heatmap. Set the radius to a sensible starting value in map units: for a city-scale map in metres, try 500 to 1000 metres; for a neighbourhood, try 100 to 300 metres. If your layer is in a geographic coordinate system such as WGS84, the radius will be in degrees, which is confusing, so reproject your layer to a projected coordinate system first, such as the local UTM zone. Set the maximum value to a number that matches your data, or leave it on automatic and adjust later. Choose a colour ramp: the built-in heat ramp, transparent to yellow to orange to red, is a good default. Click Apply and your points will melt into a smooth glowing surface. You can also set a weight field, so that a point with a value of 10 counts ten times as much as a point with a value of 1, which is useful for mapping the intensity of events rather than just their locations.

Choosing the Right Radius

The radius is the setting that makes or breaks a heat map. Too small, and you get a map of isolated blobs that tells you nothing new. Too large, and everything smears into one big warm patch with no structure. A good rule is to start with a radius that is roughly the distance you would walk to the nearest neighbouring point, then adjust by eye. For crime data in a city, 500 metres is a reasonable start. For disease cases across a region, try 5 or 10 kilometres. Always try at least three radii and compare: the right one is the one where the pattern is stable, meaning the same hotspots appear at two different radii, rather than appearing and disappearing as you change the setting. If the hotspots move when you change the radius, you do not have a pattern yet, you have an artefact.

Styling the Heat Map So People Can Read It

Put the heat map over a light basemap, such as OpenStreetMap in a muted style, so the colours stand out without fighting the background. Make the lowest values transparent, so areas with no data disappear instead of showing a pale wash. Add a clear legend that says what the colours mean and what the radius is, because a heat map without a radius is meaningless. Consider adding the original points as a faint layer underneath, so the reader can see the data behind the surface. And keep the colour ramp honest: a rainbow ramp looks pretty but misleads the eye, because the human eye does not read rainbow colours as a single scale. A single-hue or two-hue ramp, from transparent through yellow to deep red, is the honest choice, and it prints well in black and white too.

Common Mistakes to Avoid

The most common mistake is forgetting to reproject: a radius in degrees is meaningless, and a heat map in WGS84 will look stretched. The second is using a heat map for data that is not really point density: if your points are the centres of polygons, or if they represent areas rather than events, a choropleth map is the right tool instead. The third is over-interpreting small hotspots: with few points, a cluster of three can look like a hotspot when it is just chance. The fourth is hiding the data: always show the points or state the point count, so the reader knows whether the map is built on ten points or ten thousand. A heat map is a powerful tool for showing where things cluster, but it is also one of the easiest maps to make lie. Use it with a clear radius, a clear legend and a clear conscience, and it will tell the truth.