Graph and community detection analysis of pipeline network configuration and urban morphology in Kisumu and Kigali

Abstract

Due to lack of formal planning and resources, piped networks in low-income African urban areas may develop into a disorganized ‘spaghetti’ configuration. This makes such areas more vulnerable to service disruptions, with infrastructure harder to maintain and inefficiently configured. The Delegated Management Model (DMM), where a utility provider delegates water service delivery responsibilities to microoperators, is a proposed solution to this issue, implemented in numerous cities including Kisumu, Kenya. However, it is unclear empirically whether DMM rationalizes the spatial configuration of piped water networks in slums. This study aims to assess how piped network topology and thereby efficiency and interruption vulnerability vary in Kisumu and Kigali based on water service management arrangements and for slum versus non-slum areas. We applied the hierarchical ‘Infomap’ community detection
algorithm to a water pipeline network digital map layer for both cities. Infomap
iteratively distinguishes communities as areas of the network where dense connections exist, generating a hierarchical classification based on the network topology. Graph-based analysis highlighted the structural characteristics of pipeline networks shaped by the cities' urban planning and development strategies. In both cities, Infomap output differentiates low-income areas from other neighborhoods, in Kisumu further discriminating DMM neighborhoods from other parts of the pipeline network. This suggests that the topology and thereby vulnerability of the pipeline network within DMM neighborhoods demonstrate similarities, and that these are distinct between slum and non-slum areas.

More information

Identifiers

ORCID for Christopher T. Lloyd: orcid.org/0000-0001-7435-8230

ORCID for Jim Wright: orcid.org/0000-0002-8842-2181

Catalogue record

Export record

Altmetrics