Essentiality of Data Structure

Edwards (2010) Mixed-Method Approaches to Social Network Analysis

Gemma Edwards champions the mixed use of qualitative and quantitative methods with regards to social network analysis (SNA). Edward notes that quantitative SNA data can be presented in visual network maps (sociograms). Behind these seemingly incomprehensible webs of ties are the valuable underlying structure of the data, yielding measures such as ‘centrality’, ‘cores’ and ‘segregation’ (11). With respect to qualitative methods Edwards points out that participatory mapping, such as the ‘concentric circles’ approach, is an invaluable tool for qualitative SNA. In this practice the precise distance of contacts relative to the central actor is extraneous. For both examples it is the structural relationship of actors to other actors that is key.

This concept immediately reminded me of our GIScience seminar discussion on the significance of “the most famous map in the world” – the London Underground map a.k.a. the Tube map. To the dismay of novice geographers, the Tube map completely distorts the geographical layout of London. Distance-based measures of proximity do not matter to Tube passengers trying to get from Point A to Point B. Instead, spatial topology, the essential spatial arrangement of parts, is the critical factor for Tube navigation. The importance of the structural relationship of data to other data is the common grounds of SNA and GIScience.

Finding the bare essentials of data structure is not an unfamiliar concept to GIScience. This is a principle that was employed by Bonnell et al. (2013) in their application of geospatial agent-based modeling. Rather than accounting for an infinite number of parameters, the authors filtered out information that would be superfluous to addressing their research question, thereby yielding the fundamental elements that explained primate movement. In a time when the volume and flow of data is beginning to exceed the capacity of traditional statistical methods, quantitative methods (including GIScience) could learn a thing or two about essentiality.


Comments are closed.