Quantitative comparison of the distribution of densities in three Swedish cities

Typologies play a role in urban studies since a long time, but definitions are often rather abstract, ill-defined and at worst end in fixed stereotypes hiding underlying spatial complexity. Traditional typologies are focussing on separate elements, which allow for understanding crucial differences of one spatial feature in greater detail, but lack the capacity to capture the interrelation between elements. Further, they often focus on one scale level and therefore lack to acknowledge for interscalarity. Recent publications define morphological typologies based on quantitative variables, building on the seminal book \ub4Urban Space and Structures\ub4 by Martin and March, published in 1972, but using more advanced spatial analysis and statistics. These approaches contribute to the discussion of types in two ways: firstly, they define types in a precise and repeatable manner allowing for city-scale comparisons; secondly, they acknowledge cross-scale dynamics important for e.g. living qualities and economic processes where not only the local conditions are important, but also the qualities in proximity. This paper focuses on the comparison of building types in three Swedish cities, using the multi-variable and multi-scalar density definition. A statistical clustering method is used to classify cases according to their measured similarity across the scales. The results show that working with types is a fruitful way to reveal the individual identity of these types, compare cities and highlight some differences in the way the three cities are structured