Using cluster analysis to identify communities of interest for purposes of legislative redistricting: A case study of parishes in Louisiana

Abstract

The use of Geographic Information Systems (GIS) technology has been found to limit the kinds of communities that can in principle achieve political representation through redistricting because the process excludes “communities of interest” (COIs) that cannot be mapped. We argue that these limits can be overcome using empirically based variables to define a COI in conjunction with an empirically based classification system, cluster analysis. Our argument is based on a case study of a legislative redistricting case in which we examined the cultural, economic, demographic, historic, and social characteristics upon which a COI would be defined. We employed an empirically and scientifically based classification system called “cluster analysis” to determine if these two geographically-separated groups of parishes form a single COI, which, in turn, would serve as a justification for them being placed together into reorganized congressional District 5. We find that East Carroll Parish and its six neighboring parishes (Franklin, Madison, Morehouse, Richland, Tensas, and West Carroll) are in a different COI grouping than East Baton Rouge Parish. Moreover, from the COI perspective, East Baton Rouge Parish should not be included in proposed plans involving a proposed RCD5 that include East Carroll Parish and its six neighboring parishes. This finding also applies to Lafayette Parish, both in whole and in part. This finding is not only relevant because COIs are important in redistricting, but it demonstrates that the limits identified in regard to the use of GIS-based technology in identifying COIs can be surmounted.

Keywords:

• 2020 Census
• gerrymandering
• GIS
• political representation
• race

Data availability

The data underlying this article are secondary and available from the corresponding author.

Disclosure statement

No potential conflict of interest was reported by the author(s).

An ethics approval statement is not applicable because the data used are secondary and no human subjects review was required.

The authors did not use AI to generate any part of the paper.

A clinical-trial registration statement is not applicable because no human subjects were involved.

A statement regarding permission to reproduce material from other sources is not applicable because no excerpts from copyrighted works owned by third parties are included.