8-2. Key concepts

• A layer's data sources are what the user sees on the map. Its statistics sources are what the user queries — they are stored in a separate statistics array on the layer and are never drawn as a normal layer.
• In this tutorial the statistics sources are vector files (FlatGeoBuf, or GeoJSON) whose features are administrative boundaries. Each feature carries pre-computed attributes — in this case the area of each World Cover class inside that boundary.
• Because the numbers are computed in advance, the Explorer does not have to process any raster on the fly. It simply loads the vector file for the current view and reads the attributes of whichever feature the user clicks.
• Statistics render best when the layer also has categories defined: the class labels and colours from the layer's category list are used to present the per-feature breakdown.

• Each statistics source has a level. Level 0 is the coarsest set of boundaries and higher levels are progressively finer. The Explorer picks the level appropriate to the current map zoom, so a user sees countries when zoomed out and small regions when zoomed in.

  Here we use NUTS (Nomenclature of territorial units for statistics), the standard European hierarchy:

  Level	NUTS	Typical unit
  0	NUTS 0	Country
  1	NUTS 1	Major region
  2	NUTS 2	Basic region / province
  3	NUTS 3	Small region

• Levels are assigned in the order you add the sources, starting at 0. Add them coarsest first and the numbering takes care of itself.

• In the Explorer the user opens the Statistics tab and clicks a feature to see the summary for that area.

See Statistics for the full reference.