Nearest neighbour analysis
Topic: Geographical Data: Sources, Compilation and Processing · NCERT: Beyond NCERT
Meaning
Nearest neighbour analysis is a method from Clark and Evans (1954). It shows whether points on a map, usually settlements, are bunched together, scattered at random or evenly spaced. It works in three steps. First, measure the average distance from each settlement to its closest neighbour. Second, compare this with the distance you would expect if the settlements were scattered at random. Third, read the result as an index, Rn. An Rn near 0 means clustered, 1 means random, and values up to about 2.15 mean a perfectly regular (hexagonal) pattern. Today this is done inside a GIS (geographic information system).
Example
Villages on the flat Ganga plain of Uttar Pradesh are spread fairly evenly over the farmland, so they would give an Rn above 1. In a Himalayan valley, settlements crowd along the river, so they would give a much lower Rn, closer to clustered.
Don't confuse with
- Clustered or compact settlement (houses packed together inside one village): this describes the layout within a village. Nearest neighbour analysis measures the spacing between whole settlements.