A Consumer Income Predicting Model Based on Survey Data: an Analysis Using Geographically Weighted Regression (Gwr)

Eduardo Francisco, FGV-EAESP, Brazil
Peter Whigham, University of Otago, New Zealand
Francisco Aranha, FGV-EAESP, Brazil
Felipe Zambaldi, Universidade Metodista de Sao Paulo (UMESP), Brazil
Based on data of 662 households from 75 districts in the city of Sao Paulo, this paper investigates the relations between electricity consumption and household income with use of geographic weighted regressions (GWR). Findings reveal that electricity consumption is useful for characterizing household income, a frequently used proxy for purchasing power. The employed GWR were more effective to the studied task than traditional linear regressions. Also, alternatives for allocation of points were analyzed, because their exact locations were not available. The results may be useful for marketing professionals, policy makers, and credit agents who are committed to characterizing consumers socio-economically.
[ to cite ]:
Eduardo Francisco, Peter Whigham, Francisco Aranha, and Felipe Zambaldi (2008) ,"A Consumer Income Predicting Model Based on Survey Data: an Analysis Using Geographically Weighted Regression (Gwr)", in LA - Latin American Advances in Consumer Research Volume 2, eds. Claudia R. Acevedo, Jose Mauro C. Hernandez, and Tina M. Lowrey, Duluth, MN : Association for Consumer Research, Pages: 77-83.