Computers, Society, and Nature

Rundstrom (1995) GIS, Indigenous Peoples, and Epistemological Diversity

Despite Rundstrom’s argument that the “technoscience” of GIS is at odds with indigenous epistemologies, parallel concepts do emerge. Rundstrom’s account of the Tewa people’s “circles of interdependency” as a means of storing and preserving knowledge seems remarkably similar to the direction geographic information systems and data storage are heading. Spatial data infrastructures and cloud storage do not store all information in one machine, rather, information is distributed and called upon when needed. Additionally, Rundstrom conveniently fails to mention any ‘inscriptive’ indigenous mapping techniques that represent topographical and topological relationships like GIS. During the first GEOG 596 seminar this year we were introduced to several indigenous cartographic representations such as the Inuit bone carving representing the fjords of an arctic coastline and Polynesian stick charts that depict navigational routes. Rundstrom does however illustrate how maps have authoritative power and are therefore an exploitable resource.

To me the most intriguing aspect of this article is Rundstrom’s assertion that GIS does not capture relatedness but reconstructs it.  Further, he acknowledges that reassembly of phenomena from fragments is subject to current culture-specific understanding of the world. This is something to keep in mind as GIS users: the decisions we make are products of our time. Ironically, this notion may also provide some insight as to explain Rundstrom’s treatment of indigenous societies in this article. Throughout this article he refers to the indigenous conceptions of the world as if they are isolated and singular e.g. the “Tewa’s pueblo world” and “their world” in reference to the Inuit. Rundstrom simultaneously expresses the GIS serves a singular view of the world while reducing indigenous epistemologies into an ideal form. Treating entire cultures as if they are homogenous seems to discredit his own argument. This paradox reinforces the fact that cultural complexity is a difficult issue to discuss.

-BCBD

“Mixed-Method Approaches to Social Network Analysis” describes the various takes on Social Network Analysis (SNA) using both quantitative and qualitative data. Whereas quantitative data is used in traditional research and allows to “map and measure certain aspects of social relations in a systematic and precise fashion”, quantitative data brings the specific benefit of “(adding) an awareness of process, change, content and context” (Edwards, 5). Using both approaches, or mixed-methods, allows the researcher to get an ‘inside’ and an ‘outside’ view of the subject.

When reading the article, I was struck by the void in which the social networks exist. Despite using the distinctly spatial language of mapping networks, there is a lack of spatiality in the discussion. If we remember Tobler’s first law of geography, “Everything is related to everything else, but near things are more related than distant things”, distance should be included in all networks. Without weighted relationships or spatial and temporal data, the networks described in the article seemed superficial.

GIScience could take SNA to new dimension: space and (with the right Temporal GIS) time. GIS already has a network analysis function, which could be used to link personal ties between individuals.

With many social networks now only existing in the virtual world, space takes on new meanings (as we saw with spatial cyberinfrastructure). Does physical location then matter in these networks? I would argue that it does, as it is loaded with qualitative data (social and political context, intent, etc) and quantitative data (the coordinates).

-IMC

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.

-BCBD

My very first thought this week was: “Why are we being assigned this article? What does social networks analysis have to do with GIScience?” A few sentences into the article and it struck me like an anvil over the head – EVERYTHING is related to GIScience nowadays. On this note, when reading this article I was able to draw two small parallels and an overall large realization.

Firstly, in network analysis data management is obviously key. There is just so much data out there these days (within a year the average person produces 1.8 million mega bytes of data – that’s 9 CDs a day). Therefore, it is absolutely necessary to organize the data into a fashion which makes it possible to analyze. This was the easiest parallel to draw – when we get spatial data that we want to analyze in ArcMap we have to organize it first.

Secondly, there is the data analysis portion of network analysis (surprise, surprise). Of course there is the obvious: you have to run statistical models in GIS and in network analysis. This second parallel brought me to my large realization (call it an epiphany if you will): this entire article is a debate on what type of analysis is considered tangible or scientifically legitimate. This reminds me of that pesky background argument – is GIS a science or a tool? Is qualitative analysis a legitimate way of network analysis? Seem a bit familiar? It even gives the three options that we get in the GIScience debate: one (quantitative/tool), the other (qualitative/science), or mixed (a happy marriage of both). Seeing as I somewhat agree with the qualitative argument in network analysis, this got me considering the ‘s’ in GIS as a science…down the rabbit hole we go.

Until next time,

Nod

In the review paper titled ‘Mixed-Method Approaches to Social Network Analysis’, Edwards begins by outlining the two distinct approaches that one can adopt in the study of social networks through Social Network Analysis (SNA). A network can be described as a constellation of interconnected nodes linked to one another through lines that represent flows or relationships. In SNA quantitative approaches measure network properties such as density, segregation and centrality in a precise fashion, whereas qualitative approaches are more interested in the meaning of this structure, the process of how it came about and the context in which the network is found.

The two approaches affect more than just the way in which the analysis of the networks is executed, but also how the data is collected. The primary modes of data collection for quantitative analysis are walking interviews and visual mapping in which agents or nodes express their perception on the quality and nature of their links with other actors. This method of data collection brings to mind participatory GIScience, where by which the actors of the network volunteer information from an inside view which allow the study of the means and context of the network. This data can then be translated to relational data stored in adjacency matrix which stores information about which agents are tied to which and in what direction. This kind of data is more associated with quantitative analysis.

Sociograms can be described as a is a graphic representation of social links and can be used to both display network structure and offer insight to quantitative researchers who may have missed on linkages and dynamics when analyzing their ethnographic raw data. The link to GIScience demonstrated in this aspect of SNA is the significance of proximity, the position of nodes in a sociogram affect the how one may understand the network. Two unrelated nodes may be positioned close to each other without any links between them, however the mere fact that they are situated next to each other would lead one to believe they are more related to each other than nodes further away (Toblers’ law).

I would echo the push this article makes for a mixed methods approach to SNA, and I believe there are similar need in the realm of GIScience. Precise measurements without context and meaning are weak.

– Othello

The idea of Social Network Analysis (SNA) is an interesting one, something that I have seen in different areas of study, yet never really gave much thought in thinking of it as a discipline, or method of analysis. The article talks about how recently, it has been used to quantitatively analyze social networks and the interactions that occur between nodes in a mathematical approach. In the past, networks were studied in a more qualitative way. The objective of the article is to demonstrate how there is not one correct answer when it comes to SNA, but rather a complement of both quantitative and qualitative analyses in whatever degree necessary is most beneficial.
Although slightly confusing, I can understand the benefits to both types of analyses. When studying social interactions amongst a group of subjects, a pure quantitative analysis appears strong and scientifically sound. However, when you try to dig deeper into the relationships that occur within the networks, quantitative analyses only go so far in trying to paint a complete picture. It is within the qualitative analyses that you can tease out the complex details of the interactions within the network. The author does a good job in demonstrating how a mixed method approach is the most beneficial, and that the structures of networks are so complex that a single qualitative or quantitative approach is inadequate for proper analysis.
It makes a lot of sense to view a social network in this sense. For example, when studying the movements and interactions among indigenous groups in Liberia, a mathematical computer model could probably predict and explain the dynamics to a certain degree, however only with a more qualitative understanding of the situation would you be able to characterize the whole situation. The movement of SNA from a pure quantitative or qualitative state to a mixed method state appears to be a very positive thing, that will help to better understand the dynamics of complex social networks.
Buzz

Gemma Edwards’ article argues that a mix of quantitative and qualitative approaches in social network analyses would enable an understanding of both the structure and process of the network.
Social networks analysis evaluates social relationships among a set of actors, reminiscent of topological associations among objects or locations.
Adding a spatial component to social network analysis would enhance our understanding of the structure of the social network, and could provide insights into the process of social networks change over time, importing all the conceptual underpinnings geography and GIScience.
In so doing, concerns of scale and visualization become particularly important. At what unit/level should you explore relationships? At what spatial scale do you store information about the chose unit of analysis? How to visually represent movement, in a legible, informative way? When is it necessary to be spatially accurate, and when is it appropriate to forgo geography to showcase relationships? The answers to these questions will largely depend on the question being posed.
With the possibility for large spatial datasets, how do we store information? Should we use cyber-infrastructures?

Spatial Social Networks Analysis promises to deepen our understanding of how societies functions and how individuals within them relate to each other. With availability of geographically explict ‘Big Data’, the ethical, societal and political implications of such study need to be explored further.
Fan_G