ABSTRACT
Continuing the task of demonstrating the value of numeric analysis, classification of property price data is propagated as an option in this chapter. For this purpose, a dataset comprising free market transactions of approximately 46,000 dwellings during 1986–2002 was prepared by the municipal tax authorities of Amsterdam. This dataset is applied for the calculation of property taxes and contains recorded information on actual property sales prices and a number of attributes such as floor-space and subjective quality indicators. The latter are constructed by the assessor on an ordinal 10-point scale relating to the house, its maintenance and the micro-location. Marks 1 (very bad) to 10 (perfect) are given for quality, maintenance and situation (0 is empty or unknown), respectively. This dataset can be conveniently disaggregated at district or house type levels.
