Topology errors are the invisible cracks in vector data: polygons that overlap, gaps between neighbours that should share a border, dangling lines that stop just short of an intersection, and duplicate points stacked on top of each other. A map can look perfect while its data fails every area calculation, because a parcel counted twice inflates a district total by exactly the duplicated area, and a gap between two zones leaves land unassigned in any overlay. Cleaning topology is therefore not cosmetic work; it is what makes GIS analysis trustworthy.

The most common errors each have a name and a fix. Overlaps occur when two polygons claim the same ground, often after merging datasets from different sources. Gaps, or slivers, are narrow polygons left between neighbours when boundaries were digitised separately. Dangles are lines that overshoot or undershoot their target by a few centimetres or metres, common in road networks traced from imagery. Duplicate geometries arise when an import runs twice or a copy-paste goes wrong. A district boundary file with 300 polygons may easily contain 40 overlaps and 25 slivers after a careless merge, and each one will distort area summaries until removed.

A final check that catches many errors is to run a simple overlay after cleaning: intersect your cleaned layer with itself and confirm the result contains no polygons, and run a dissolved area summary by district to confirm totals match official figures within one percent. If a district total jumps by three percent after cleaning, stop and investigate rather than publishing, because that jump usually means a real boundary was snapped to the wrong neighbour or a large sliver was deleted that was actually a genuine enclave. Cleaning should make totals more accurate, not merely smaller, and a documented before-and-after table is the evidence that it did.