ABSTRACT

This chapter explains how to take spatial microdata, and allocate the resulting individuals into household units. When the individual level data are independent from the household level data, they can rarely be linked. Data coming from different sources, sometimes implying different total populations, can cause this inconsistency. The chapter describes the method proposed by Johan Barthelemy for dealing with such situations. The household generation process ends after any one of these three conditions: if all households have been constructed; if the pool of individual is empty; or if the process fails to find a member for a household in the previously generated ones. In some cases, information about households is included in the individual dataset. The assembling of children to the head has been made by a similar process and gives as good results in terms of distribution of ages.