Relatively few quantifiable and objective methods exist for the analysis of architectural elevations. Only one of these methods has been repeated by multiple researchers and used for the analysis of a wide range of historic and modern buildings and architectural types. In the present paper, the computational variation of this method–the fractal approach to determining characteristic visual complexity–is applied to the early house designs of Peter Eisenman. The results of this analysis are then subjected, for the first time, to cluster analysis in an attempt to uncover patterns in the way in which Eisenman’s houses are shaped by orientation, address and permeability. Such an analysis is of interest because Eisenman argues that he set out to design these early houses without regard for such factors. The paper concludes that the mathematical data generally supports Eisenman’s contentions regarding his Houses I, II, III, IV and VI.
27th Conference on Education and Research in Computer Aided Architectural Design in Europe (eCAADe 2009). Computation: The New Realm of Architectural Design: Proceedings of the 27th Conference on Education and Research in Computer Aided Architectural Design in Europe (Istanbul, Turkey 16-19 September, 2009) p. 729-735