Combining cluster analysis and GIS maps to characterise building stock: case study in the historical city centre of Viseu, Portugal
Catarina P. Mouraz, Ricardo M. S. F. Almeida, J. Mendes Silva
Abstract
The first step towards rehabilitation actions is deep knowledge of buildings characteristics and their conservation state. Studies comprehending large groups of buildings can be useful when dealing with significant amount of data, positively influencing decision-making processes starting from low level information. This paper describes a methodology that makes use of GIS-based maps to visualise the results of cluster analysis regarding defects in 495 buildings. This information was organised to apply a clustering technique with the purpose of identifying patterns related with defects. The GIS-based maps were created to support the interpretation of the clustering results. Conclusions regarding the most interesting number of clusters were drawn, as well as insights on the correlation between the spatial distribution of buildings and the defects observed. In two of the four analyses conducted, a pattern of buildings located in the central area of the historical zone was identified, which was associated with a higher number of defects probably related with the age of the constructions. These results are particularly relevant considering the definition of future rehabilitation actions and confirm the advantage of combining both tools, which is particularly meaningful in contexts where methodologies like this are scarce and limited in number.