Most joins in GIS match records by a shared code: a district ID in one table matching a district ID in another. But real-world questions are often spatial, not administrative. How many boreholes fall inside each watershed? Which shops are within each sales territory? Those questions need a spatial join, which combines layers based on where features are, not on what they are called.

A spatial join has a target layer and a join layer, and the direction matters. Joining points to polygons, for example schools to districts, attaches each district's attributes to the schools inside it, or counts the schools per district. Joining polygons to points does the reverse. Before you run anything, state your question in one sentence: 'for each district, count the schools inside it.' That sentence tells you the target, the join layer, and the summary you need.

In QGIS, the Join Attributes by Location tool handles this. You pick the target and join layers, choose the geometric predicate, usually 'within' or 'intersects', and decide whether to keep all target features or only those with a match, and whether to summarise or list every match. The choices sound technical, but each one maps directly to a plain-language decision about your question.