The same point data can look like a few big clusters or many small ones depending only on zoom level

What it is

Interactive maps typically tie their aggregation parameters to zoom level: zoomed out, a large radius yields a few big clusters; zoomed in, a small radius splits them into many. The catch is that the same point data can show a different number of clusters, in different places, at different strengths across zoom levels, and readers reach different conclusions at each level. The visible pattern is the product of "data x current aggregation parameters", and readers are usually aware only of the first factor.

Why it happens

The aggregation radius is the modifiable areal unit problem (MAUP) running in real time on an interactive map. The radius decides which points group together: changing it regroups assignments, shifts every cell's count and location, and reshapes the relative strength of clusters. Two adjacent subclusters at one scale merge into a single large cluster at a coarser scale; an independent hotspot at one scale may decompose into noise at a finer one. The existence and shape of the pattern are partly parameter-driven. A reader pausing at different zoom levels is silently switching between aggregation schemes, and the map does not flag it—zooming feels like a pure viewing action, when it actually changes the analysis scale.

Where it stops holding

Not every conclusion is scale-bound: global totals and the overall extent of the point set are insensitive to aggregation parameters. When the main pattern holds across multiple scales, the conclusion is scale-robust and the parameter question recedes. What is affected are conclusions that depend on cluster structure itself ("how many hotspots", "where are they") and, above all, comparisons made between zoom levels—those may not even compare the same set of units.

Applying it

• Before drawing a cluster-based conclusion, recheck the pattern at two or three zoom levels; conclusions drawn at a single level should carry their aggregation parameters.
• When exporting static maps or reporting screenshots, fix the zoom level and radius, and state both in the caption so readers do not treat the view as the data's inherent form.
• Verify with a scale-sensitivity test: rerun the same analysis at multiple scales and claim robustness only when the conclusion holds across all of them.

Related

• Same group: U12.03.1 Point maps systematically undercount in high-density regions because of overplotting · U12.03.2 Aggregation into clusters hides within-cluster distribution differences · U12.03.3 Cluster markers must encode how many points they contain · U12.03.4 Hexagonal binning or density surfaces can replace direct point overlay
• Nearby: U12.03.2 Aggregation into clusters hides within-cluster distribution differences · U8.06.2 The choice of grouping can flip a conclusion's direction
• Search terms: modifiable areal unit problem · scale dependence · zoom-dependent aggregation