# Aggregation

The structures, roads, sidewalks, etc. in a college campus are aggregated into a single "campus" object.

Aggregation is the practice of grouping multiple objects with similar characteristics into a single object, thus creating a layer with less detail than the original data.[1] This makes aggregation a useful tool in cartographic generalization, as it simplifies the data so that it is easier to draw and interpret. For example, in census data parcels are aggregated into districts (e.g., blocks, tracts, counties) to make the distributed data easier to analyze and to protect the privacy of individuals.[2] Aggregation differs from the Merge tool by creating a higher level of dimensionality (point-to-line, point-to-polygon, line-to-polygon).

## Theory

This image is of a highway passing through a city. The points are accidents happening on or near (at entrances) the highway in the month of June.
This is the same image after an aggregate operator was performed. The points are changed into polygons, showing areas where accidents occurred in June rather than individual points and increasing dimensionality.

As in any form of generalization, aggregation eliminates a certain amount of detail (both spatial and attribute) but creates simplicity for the user who is more interested in the unit as a whole and not each individual part within the whole. This could be the case if the gatherer of data intends to preserve the privacy of the individuals sampled, or if the raw data collected is too large or too complex to accurately and conveniently convey the information on a map.

Thematic maps depend on aggregation in order to convey the purpose of the map. However, because the individuals included in the aggregate are rarely homogeneous, care should be taken to avoid the Ecological fallacy, which is assuming that each individual within the given area possesses the attributes represented by the aggregation. Returning to the census example, if a particular neighborhood has a high Mean Annual Household income, it should not be assumed that each individual in the area has a high Annual Household Income, as the neighborhood could include many households above or below that income level (with an unknown standard deviation).