How to buffer features and count what falls inside

The short answer: buffer the source features by your distance, then select the target features that intersect the buffer, then read the count. Three steps, in every tool that has ever existed.

"How many X are within Y metres of Z" is the question GIS gets asked more than any other. Everything else is a variation.

Getting the question right first

Two things decide whether your answer is right, and neither is about software.

"Within 500 m" of what part? Of the centre of each parcel, or of its boundary? A 10-hectare block whose edge is 400 m away but whose centroid is 900 m away is either in or out depending on an answer nobody stated. Buffering the polygons and intersecting handles this properly: measuring centroid to centroid does not.

Intersects, or is contained by? A parcel half inside the buffer intersects it, but is not within it. These give different numbers, and the difference is often large. Decide which one your question means before you run anything, because both are defensible and only one is what you were asked.

In QGIS

1. Vector, Geoprocessing Tools, Buffer. Input is your source layer, distance is your number, and check Dissolve result if overlapping buffers should merge into one, which they usually should for counting.
2. Vector, Research Tools, Select by Location. Select from your target layer, where features intersect, comparing to the buffer layer.
3. Read the count off the layer panel, or open the attribute table and look at the selected total.

The trap: if your layer is in degrees (EPSG:4326), a buffer distance of 500 means 500 degrees, which is nonsense. QGIS may not stop you. Reproject to a metric CRS first, or use the buffer tool's ellipsoidal measurement option. This is the single most common way this analysis goes silently wrong.

In the browser

Same three steps, no install. In the Geo Hound workbench:

1. Select the source features. Buffer works on the selection, so use Identify, Box, or By attribute to pick them first.
2. Open the Buffer tool, set the distance and unit (m, km, ft, mi), and click Buffer selection. You get a new result layer.
3. Open By location, select from your target layer, intersecting the buffer layer, and click Select by location. The selection status tells you how many.

The dock opens with the selected features in a table, and CSV exports them.

Units are metric or imperial explicitly, so the degrees trap does not arise.

By asking

The version that needs neither the tool names nor the sequence. In the assistant:

"Buffer the wells by 500 metres and tell me how many parcels fall inside."

It runs the same three operations on the same geometry and gives you the number, the buffer as a layer, and the parcels selected on the map so you can look at them. One credit, three operations.

The reason to trust it is that it is not the model doing arithmetic: it is the model choosing to call buffer, then select-by-location, then count, and the geometry library doing the work. More on why that distinction matters.

Checking your answer

Whichever way you did it, before you report the number:

Look at the map. Buffers are the easiest thing in GIS to eyeball. If the rings look too big or too small against a scale bar, they are.

Check you have all the features. This is the one that ruins analyses quietly. Web services cap what they return, typically at 1000 or 2000 features. If you loaded a layer from a service and it only fetched your current view, you just counted the parcels in your current view. In Geo Hound, that is what the Fetch all button on a layer card is for; in QGIS, export the layer to a GeoPackage first.

A count that comes back as exactly 1000 is not a coincidence. It is a truncation.

Check the CRS. If the number is wildly off, this or the point above is why, roughly always.

Ask whether overlapping buffers double-counted. If two wells are 200 m apart and you buffered each by 500 m without dissolving, a parcel between them may be counted twice depending on how you counted. Dissolve, or count distinct targets rather than intersections.

Where to go from here

Buffer-and-count is the first move. The follow-up is usually a spatial join, to get which ones and their attributes rather than just how many, or a nearest-distance to ask how far rather than how many.

If the data you want to do this to is sitting in a public web map right now, Geo Hound catches the service behind it while you browse, so you can go from looking at somebody's map to having buffered its data in about a minute.