@inproceedings{discovery1472823,
            note = {This version is the version of record. For information on re-use, please refer to the publisher's terms and conditions.
Copyright {\copyright} Space Syntax Laboratory, UCL, 2015.},
         journal = {Proceedings of the 10th International Space Syntax Symposium},
       publisher = {Space Syntax Laboratory, The Bartlett School of Architecture, University College London},
         address = {London, United Kingdom},
           title = {Characterising Place by Scene Depth},
       booktitle = {Proceedings of the 10th Space Syntax Symposium (SSS10)},
          series = {International Space Syntax Symposium},
          volume = {10},
           month = {July},
           pages = {120:1--120:15},
            year = {2015},
          editor = {K Karimi and L Vaughan and K Sailer and G Palaiologou and T Bolton},
        keywords = {Isovist, place, classification.},
             url = {http://www.sss10.bartlett.ucl.ac.uk/category/07-environmental-and-spatial-cognition/},
          author = {Davis, A and Hanna, S and Aish, F},
        abstract = {Turner and Penn introduced the notion of integration of isovist fields as a means to understand such fields syntactically - as a set of components with a structural
relationship to a global whole (1999). This research was further refined to put forward the concept of visibility graph analysis (VGA) as a tool for architectural analysis (Turner, Doxa, O'sullivan, \& Penn, 2001), which has become widely used. We suggest a complementary method of characterising place that does not make use of integration or a graph yet which allows - as visibility graph analysis does - discrete view points to be dimensioned in relation to a set of such viewpoints. In our method, Principal Component Analysis (PCA), a statistical technique, is employed to infer salient characteristics of a set of views and then to situate these component views within a low dimensional space in order to compare the extent to which each
view corresponds to these characteristics. We demonstrate the method by reference to two distinct urban areas with differing spatial characteristics. Because PCA
operates on vectors, order of the data has important implications. We consider some of these implications including view orientation and chirality (handedness) and
assess the variance of results with regard to these factors.}
}