Need for an Integrated Deprived Area “Slum” Mapping System (IDeAMapS) in LMICs

Abstract

Ninety percent of the people added to the planet over the next 30 years will live in African and Asian cities, and a large portion of these populations will reside in deprived neighborhoods defined by slum conditions, informal settlement, or inadequate housing. The four current approaches to neighborhood deprivation mapping are largely silo-ed, and each fall short of producing accurate, timely, comparable maps that reflect local contexts. The first approach, classifying "slum households" in census and survey data and aggregating to administrative areas, reflects household-level rather than neighborhood-level deprivation. The second approach, fieldbased mapping, can produce the most accurate and context-relevant maps for a given neighborhood, however it requires substantial resources, preventing up-scaling. The third and fourth approaches, human interpretation and machine classification of satellite, aerial, or drone imagery, both overemphasize informal settlements, and fail to represent key social characteristics of deprived areas such as lack of tenure, exposure to pollution, and lack of basic public services. The latter, machine classification of imagery, can be automated and extended to incorporate new and multiple sources of data. We summarize common areas of understanding, and present a set of requirements and a framework to produce routine, accurate maps of deprived urban areas that can be used by local-to-international stakeholders for advocacy, planning, and decision-making.

Exploring convolutional neural networks and spatial video for on-the-ground mapping in informal settlements

“…Only at this scale, with these types of data, can an accurate assessment of the interaction between living conditions and potential environmental health risks be identified. To use machine learning at this scale, to capture factors that often occur beneath the overlapping building canopy and therefore beyond normal remotely sensed imagery [45], a new image library is required. These data also need to be longitudinal given the dynamic nature of these spaces, with significant changes occurring at different cadences, both seasonally and then from year to year [9,13].…”

“…There is no easy solution to solve this gap; online visual data suitable for automated image classification in informal settlements is scarce, especially when the additional problem of how these environments change geographically; similar settlements in Haiti and Ghana have similar problems and features, but the details needed for image classification vary considerably. While remotely sensed imagery can be improved with other data sources [18] such as local censuses, there is still a need to contextualize local environment at the street scale [45] with onthe-ground imagery to improve the generalization and accuracy of machine learning models. While normally collecting these types of data are logistically challenging, the project team for this study has been using SV in multiple environments and time periods, amassing a considerable library of granular environmental imagery which can be used to explore various aspects of model training for these types of settings.…”

A Global Estimate of the Size and Location of Informal Settlements

“…UN-Habitat calculates slum estimates from household surveys, such as demographic and health surveys, multiple-indicator cluster surveys, or national survey initiatives. Recent studies have revealed that the current sources used to quantify slums are structurally underestimating the slum population, as sample sizes are often based on outdated census data and smaller slum pockets are generally difficult include in current data collection methods [11]. Another concern is in relation to the available data format.…”

“…It assesses the multiple deprivations of slums by combining different data collection methods (censuses, surveys, field-based mapping, human image interpretation, and machine image classification) to provide a more comprehensive picture of slum realities. Punctual case study research has revealed a wide gap between deprived areas and officially recognized slums [11]. The upscaling of this framework can potentially address most of the identified shortcomings of current data collection [9].…”

Gridded Population Survey Sampling: A Review of the Field and Strategic Research Agenda

“…Also, as urban settlement classification becomes increasingly possible [57], survey designers need to understand how within-urban stratification affects the various sample designs used in gridded population, and other, surveys. With no way to stratify urban populations, all surveys are at risk of under-sampling or omitting slums and other vulnerable populations [58,59]. In addition, research is needed to balance survey designs that can support both precise design-based estimation of outcomes and precise SAEs of indicators at fine geographic scales [60].…”