In many domains of application, sampling surveys allow to make important savings by reducing the number of persons to interview, the number of samples to be analyzed, the number of sites to be visited, etc.

If suitably organized (that is by respecting the rules of the statistics), a sampling survey can provide results as good or even better (in terms of precision) that a systematic inquiry.

In particular, the use of optimized stratified sampling and multiple level sampling (a.k.a. "cluster sampling") are very effective, if some conditions are verified.

The combined use of satellite imagery (to segment the space in homogeneous regions defining the strata of the sample), of sampled aerial photos as a basis for clusters, and of field survey to measure variables proved to be extremely efficient in many cases. The potential of spatial analysis offered by Geographical Information Systems is a perfect tool to prepare the sampling scheme, but also to further analyse the results.

We used these methods successfully for numerous estimations of population in urban zone, socio-demographic inquiries in rural zones, analysis of vegetation and health inquiries.