ST_SetEffectiveArea

What is ST_SetEffectiveArea?

ST_SetEffectiveArea is a PostGIS function that computes the Visvalingam-Whyatt effective area for each vertex of a geometry and stores it in the vertex's M coordinate. Vertices with larger effective areas are more significant for shape — those with small values are the first candidates for simplification.

ST_SetEffectiveArea(geometry geom, float threshold = 0, integer set_area = 1) → geometry

With threshold = 0, every vertex gets an M-value and no vertex is dropped. With a positive threshold, vertices below that threshold are removed.

When would you use ST_SetEffectiveArea?

Use ST_SetEffectiveArea when you want progressive simplification — storing effective areas once so that multiple simplification levels can be derived cheaply by filtering on M-values. It powers zoom-dependent rendering in vector-tile pipelines where each zoom level shows progressively more vertices.

1SELECT id, ST_FilterByM(ST_SetEffectiveArea(geom), :zoom_threshold) AS tile_geom
2FROM features;

FAQs

How is this different from ST_SimplifyVW?

ST_SimplifyVW applies the Visvalingam-Whyatt algorithm and drops vertices below a threshold in one pass. ST_SetEffectiveArea computes and stores the per-vertex effective area without dropping, so you can decide later — or at multiple thresholds — which vertices to keep. The two are typically used together via ST_FilterByM.

What are the units of the effective area?

Squared units of the input's SRID. For UTM that is m²; for EPSG:4326 that is degree². For meaningful, comparable thresholds, reproject to a metric CRS before calling.

Does it overwrite existing M-values?

Yes, by default. If your geometry already carries M-values you want to keep (e.g. timestamps), consider calling ST_SetEffectiveArea on a copy and storing the results separately.

Why is this useful for vector tiles?

At different zoom levels you want different levels of detail. By pre-computing effective areas once, you can serve each zoom level by filtering vertices whose M-value exceeds a zoom-dependent threshold — much cheaper than running a full simplification per tile.