The Six Cities Project: developing a methodology of surveying densely populated areas using social science assisted and diachronic remote sensing based classification of habitation

Abstract

This paper provides a statistical evaluation of the methodology of the NSF funded Six Cities Project. The project develops a methodology for surveying densely inhabited areas by processing diachronic remote sensing imagery to create habitation strata or urban classes. These classes become part of a sampling strategy which gives every pixel associated with habitation a specified chance of selection and then draws a representative sample of pixels. These pixels become center points for household surveys which can study a variety of issues including health, environment, livelihood strategies, demographics and household labor, expenditures and income. The methodology lends itself to GIS construction and the generation of data that can be easily compared and can be of maximal use to municipalities, governments, scholars and NGOs. It also provides a long term basis for inexpensive surveys that can have a high claim to reliability and representativity.

Key words: remote sensing, urbanism, survey methodology, National Science Foundation, health, environment, livelihood strategies, demographic, household labor, expenditures, income, Africa, Middle East, Morocco, Senegal, Mali, Niger, Tanzania, Botswana, Marrakech, Dakar, Bamako, Niamey, Dodoma,  Gaborone.