Topology in GIS describes the spatial relationships among features — adjacency, connectivity and containment — independent of exact coordinates. Topology rules encode integrity constraints: parcels must not overlap, pipe networks must connect at nodes, administrative boundaries must share edges without slivers. Data that violates these rules produces wrong areas, broken network traces and analysis artefacts, which is why topology validation is central to professional data production.
The relationships topology captures
Connectivity says which lines meet at shared nodes — the foundation of network tracing in roads and utilities. Adjacency says which polygons share boundaries, enabling neighbourhood analysis. Containment says what lies inside what, from buildings in parcels to islands in lakes.
These relationships can be stored explicitly, as in classic coverage models and modern network datasets, or enforced as validation rules over simple geometries, the approach most contemporary geodatabases take.
Classic topology errors
Slivers are needle-thin polygons born where two boundaries that should coincide almost do. Gaps and overlaps break the rule that a parcel fabric must tile space exactly once. Undershoots and overshoots leave line ends dangling short of, or past, their intended junctions, severing network connectivity.
Each error type has a signature cause — digitising without snapping, merging data from different scales, careless format conversion — and each silently corrupts downstream analysis until systematically found and repaired.
Building topology-clean data
Production discipline beats after-the-fact repair: snapping environments, shared-edge editing, and topology rules validated continuously during capture. When repairing legacy data, cluster tolerances and rule-based error inventories turn an amorphous mess into a burn-down list.
Quality standards for cadastral, utility and basemap data routinely specify permissible topology error rates near zero — achieving that at scale is one of the clearest dividers between hobby digitisation and production-grade GIS data services.