In this paper, I try to apply the logic of Space Syntax (sometime I might write something explaining what this is, if you are interested, and if you have access to an academic library, this is a good introduction; if you want to read the book on it, one of the major works is available online here. It’s a big file, but it downloads very quickly on my computer).
Anyway, syntactical analysis is a method of analyzing a network, generally of lines-of-sight or isovists, that has been shown to be rather reliable in predicting areas where pedestrians and automobiles will gather. I am planning on using it extensively in my MP.
The real chore in this project was finding a program that would conduct spatial analysis using nodes instead of lines. A program I am using in my MP is Depthmap(1), which has output that looks like this (part of an urban design project this year – a greenfield site that is based on Loket, Czech Republic(Google Maps) :
As you can see, each line has a value based on its connections to other lines. Most syntactic software works in this manner. After much searching, I found AGRAPH(2), which allows one to make a network of nodes, such as subway stations. Here is what that looks like in looking at the Washington Metro:
The program is somewhat unwieldy, as the nodes cannot be made smaller except by using a computer screen with a higher resolution. Therefore, the placing of the nodes becomes more difficult when you want to place many of them, as in the Paris metro:
Anyway, I graphed nineteen metro systems (3) into AGRAPH, and looked to see whether one particular syntactical measurement, intelligibility, is correlated to boardings/population. This regression (4) did not give significant results, but this may be due to the small size of the sample. The scatter plot does suggest that there might be a relationship, and a larger study could take place. Second, I looked at several syntactical measurements, connectivity, integration, and total depth, and their correlation to boardings in the test case of the Washington Metro. The regression results here do show significant relationships between the measurements and boardings.
Several next steps are suggested by this project. First, more systems could be mapped such that a relationship between integration and ridership could be tested between systems. Second, the Washington analysis could be repeated for other systems to see if it holds true elsewhere. Finally, a larger multiple regression could be run, looking at many more factors.
(1) Turner, A. (1998). UCL Depthmap, in University College London Bartlett School of Graduate Studies. Retrieved August 17, 2009 from http://www.vr.ucl.ac.uk/depthmap/; http://www.spacesyntax.org/software/depthmap.asp.
(2) Manum, B., Rusten, E. & Benze, P. (2005) AGRAPH, Software for Drawing and Calculating Space Syntax “Node-Graphs” and Space Syntax “Axial-Maps”, in Norwegian University of Science and Technology. Retrieved December 3, 2009 from http://www.ntnu.no/ab/spacesyntax.
Manum, B., Rusten, E. & Benze, P. (n.d.) AGRAPH, Software for Drawing and Calculating Space Syntax “Node-Graphs” and Space Syntax “Axial-Maps”, in Norwegian University of Science and Technology. Retrieved December 5, 2009 from http://www.spacesyntax.tudelft.nl/media/Long%20papers%20I/agraph.pdf.
(3) Choice of systems and data on ridership thanks to: Derrible, S. & Kennedy, K. (2009). A network analysis of subway systems in the world using updated graph theory. Transportation Research Board 2009 Annual Meeting.
(4) Thanks to: Tab Combs, Amanda Dwelley, Ben Owen and Eric Schultheis for help with reading the Stata output.
