Qualitative GIS

I’m having way too much trouble with a paper I’m writing. Been suffering writer’s block for a couple weeks now, and really need to start pushing stuff out. So I’m going to write the section on qualitative geographic information systems (QGIS) here, hoping that the decreased pressure will help the creative juices flow. Context: this will be a section of a paper I’m writing on my thesis work, targeted for an academic journal (Cartographica). If any of you have thoughts (content, writing style, line of thinking, etc.), I’d love to hear them. Please note that this is an incomplete first draft of one section of the paper. It is by no means complete.

Efforts to incorporate personal voices, impressions, and emotions in a predominantly computational environment find immediate resonances in qualitative geographic information systems (QGIS). Although geographers have conducted mixed-methods research for decades, the representation and analysis of qualitative data within geographic information systems (GIS) has only recently become a major concern of researchers and practitioners. ‘Qualitative’ here is defined not simply as non-numeric, but also richly contextual, personally experienced, and expressed in nuanced personal terms (Elwood and Cope 2009). Partly responding to early critiques of GIS as a quantitative, positivist, and corporatist technology, researchers have developed ways of opening GIS to alternative epistemologies and situated knowledges (Sheppard 2005). These are constituted by the everyday experiences and knowledges through which people approach and utilize technologies. In this way the genealogy of QGIS can be drawn to the critical GIS debates of the 1990s (Wilson 2009) and feminist GIS (e.g., Kwan 2002). For all the diversity in particular approaches to this engagement with GIS, the overarching corpus of work seeks to position qualitative ways of knowing and experiencing places alongside the numeric summative attributes of places. Jung and Elwood (2009) have argued that QGIS research can typically be characterized in one of three ways: (1) transforming qualitative data into a form easily represented cartographically, (2) hyperlinking from GIS-based objects to external qualitative data artifacts, and (3) modifying existing technologies on the software level to more readily handle qualitative data. Extending GIS in these ways is thought to benefit from a combination of qualitative and quantitative forms of reasoning. Whereas GIS is commonly - if not mistakenly - understood to be best suited for quantitative representation and analysis, some working with qualitative data have productively pointed to ways GIS can be conceptualized as a ‘qualifying’ and qualitatively-enabled technology (Pain et al. 2006; Kwan 2002). These lines of reasoning seek to show on the one hand how quantification can obscure the contexts and nuance through which people experience their environments, but on the other the ways that GIS can embed quantitative data with this richness. The importance of explicitly integrating qualitative or “qualified” data with GIS stems from at least three concerns. First, reflecting longstanding methodological debates, qualitative data is seen as preserving much of the rich contextual information that may be lost in a purely quantitative study (insert qualitative methods citations). In a community planning context in Humboldt Park, Chicago, Elwood (2009) effectively used qualitative data to visualize the tension between residents’ perceptions of their neighborhood and official administrative neighborhoods. Elsewhere, Elwood (2006 (Annals article)) has shown that even where quantitative data is mapped, it often must be understood in terms that lie outside the map. GIS-based mapping of vacant lots in Humboldt Park, for instance, could be understood as needs narratives, a conclusion understood only in terms of the supplementary interview data collected. A second impetus for the incorporation of qualitative data in GIS stems from the inability of much research-generated data to be represented quantitatively. Since different forms of data may represent distinct epistemological approaches to knowledge, different approaches to knowledge representation, analysis, and synthesis may lead to more productive scholarship (Lawson 1995; Pavolvskaya 2006). The third impetus lies in the political potential of integrating qualitative data in GIS. Feminist GIS in particular has emphasized the many different place-based experiences possible across different social groups. Qualitative GIS, through its combination of quantitative and qualitative representations of these experiences, is most able to capture the ways in which experiences can be represented. As one example, Knigge and Cope describe how grounded visualization, through its juxtaposition of quantitative and qualitative data, can “build on the robust capacities of ethnographic and GIS techniques to … produce rigorous results and present them in ways that are not oppressive or overly exploitive” (Knigge & Cope 2006, 2035). Further, QGIS has the potential to dislodge many of the ontological assumptions Importantly, however, QGIS need not be thought of as simply a research program; rather, many engagements with GIS already show the distinct characteristics we use to describe QGIS, in that often they work to incorporate alternative forms of representation to meet various needs. These representations may include images, sketches, perceptions, and hypothetical scenarios (Al-Kodmany 2000; Matthews 2005). If the discussion above seems to inadvertently construct an artificial binary relationship between quantitative and qualitative methods, it is worth noting that many within GIS have questioned the extent to which geographic technologies such as GIS are inherently quantitative or qualitative (Schuurman 2000). Much work in critical GIS has sought to break down this way of thinking, showing the personal narratives and knowledges that can be represented in GIS (Kwan 2002). Indeed, as Pavolvskaya (2006) has argued, rather than distinguishing between quantitative and qualitative “toolboxes”, it is often more realistic and analytically productive to think of a continuum along which most researchers operate - and along which it is possible to operate.