Why Topology Matters in Predicting Human Activities
Topology reminds us of its importance, once again. Although I’m *still* trying to actually understand it...
"This finding further supports our understanding that any spatial cognition derived from space syntax or the topological analysis is, at an aggregate level, determined by space rather than by people."
Time: August 29, 2018
Place: Morning commute (seems to feel familiar)
Pointer from: Twittery Badittery
Note type: Direct
Thus, the topology of streets differs fundamentally from the concept of topology in geographic information systems (GIS), since the latter is imposed among geometric primitives of points, lines, polygons, and pixels (Longley et al. 2015). The topology of streets differs also fundamentally from the topology in segment analysis, which is defined among adjacent line segments (Hillier and Iida 2005); see more details in Section 3. Thus, the topology of streets is unique, being a notion that is closely related to the scaling structure of far more smalls than larges
It is this scaling or fractal or living structure that makes human activities predictable
The present paper intends to illustrate that natural streets are the best topological representation for predicting human activities
The contribution of this paper is three-fold. First, we illustrate with empirical evidence that human activities are mainly shaped by the underlying scaling of far more less-connected streets than well-connected ones, and that topological analysis capable of capturing the underlying scaling can predict human activities well. Second, we demonstrate that segment analysis or conventional GIS representations are essentially geometric and therefore cannot reveal the scaling or fractal or living structure, which means they are unable to predict human activities. Third, we further clarify the importance of topological analysis in space syntax
Natural streets demonstrate far better scaling pattern than axial lines
The lack of geometric or metric information gives space syntax it's power to see the underlying scaling.
Topological analysis is capable of capturing a majority of human activities without considering metric information like street lengths and widths, and even building heights.
Geographic objects in the vector format can be identified from the imagery, but the topological relationship among the points, lines, and polygons still fails to uncover the underlying scaling hierarchy of far more small things than large ones. Therefore, as a critique to current geospatial models such as vector and raster, the geometric primitives of points, lines, polygons and pixels are mechanistically imposed, which means they pose little meanings in our perception and cognition. On the contrary, both natural streets and axial lines represent meaningful units in our cognition, and they are essentially small spaces that can be perceived form single viewpoints. Thus, the street-based topological analysis differs fundamentally from these GIS representations. In addition, GIS representations make no difference between space where people can move freely around and space occupied by buildings or other geographic features. The topological analysis concentrates only on spaces between buildings or, equivalently, streets in which people can freely move around. It is essentially the scaling hierarchy of the free space that shapes the human movement, or makes traffic predictable
To this point, we can remark that human activities as a whole cannot be well understood from individual human beings. This is because the whole, as a complex system, is more than the sum of its parts. To borrow the famous statement by Winston Churchill – that “we shape our buildings, an thereafter they shape us" – we have demonstrated through the above case studies that human activities are substantially shaped by the underlying scaling of space. It is the underlying scaling structure that makes human activities predictable. However, this predictability is at a collective scale in terms of how many people come to individual streets, rather than detailed travel routes of individual people. Researchers tend to confuse the collective traffic flow shaped by the underlying street structure with individual travel behavior. This is also a common misunderstanding in the space syntax literature, as pointed out by Omer and Jiang (2015). In other words, the traffic flow captured by axial-based space syntax has little to do with human beings or random walkers. In a given street network, up to 80 percent of traffic flow can be accounted for by the underlying street structure (e.g., Jiang and Jia 2012). This social physical perspective (Buchanan 2007) is what underlies the topological representation and analysis.
This finding further supports our understanding that any spatial cognition derived from space syntax or the topological analysis is, at an aggregate level, determined by space rather than by people
August 30th, 2018