In my previous blog post I referred to a specific South African digital “context” in which I would be exploring the emergence of smart city discourse among South African municipal governance structures. The previous blog post also described why, at the level of methodological approach, certain modes of web enquiry may not be appropriate.
This second digital methods blog post will attempt to elucidate some aspects of the “context” I previously referred to, providing insights into its form and making a case for an exceptional form that requires a more considered methodological approach.
The first key aspect of this South African governmental digital context that I encountered was the inconsistent manner and logic in which information and documentation is ordered, presented, or generally made available. Government websites across the numerous provincial and municipal levels each store and present information in a manner that is not only starkly different between cities in the country, but also between the provincial government and its corresponding major metropolitan area. This is most interesting when considering the constitutionally mandated relationships between the city and province, and how this relationship manifests in motion.
When considering digital enquiry methodologies described in the previous post such as web scraping, this inconsistency is a key hurdle. This “varied” topography makes running automated, scripted scraping programmes between various domains quite a considerably complex task. A web scraper – both of the coded and packaged variety – would require specific retooling and manual guidance in order to be “trained” for each domain.
This technical fiddly-ness is perhaps more complexity than what is suited for the type of information we are looking to obtain from websites; simply pulling relevant textual and visual data without much concern at these early stages for the more intricate operational data “behind the curtains” of the domain.
Another a key issue emerging from the South African governmental digital context is that information (both generally, and information specifically pertaining to smart cities) is not necessarily presented in an easily obtainable form for scraping tools. Specific information is often presented not as textual data on webpages, but rather in various documents embedded within webpages.
Taking smart cities for example, many South African governmental websites will not present information on their intended plans or policies as accessible information webpages to the public. Rather, such plans and intent are embedded in specific sections of urban spatial and economic development planning documents; documents that are not easily digestible or accessible by the general public. In essence, in the public-facing South African government discourse, reference to “smartification” and “smart city” planning forms a component of an overall strategy rather than a coherent, well-presented and easily-accessible strategy in of itself.
There are documents from certain municipalities and provincial governments that do present planned “smartification” action as consolidated, easily digestible presentations. These documents cover the basis of planned policy and programme action around a planned smart city discourse, detailing how plans have grown and shifted over time. These documents, however, are unfortunately often presented to particular actors (including business collectives, public private partnerships, international investors or donor agencies, but excluding the general public) with which government bodies attempt to build networks and support for their plans.
This same collection and succinct presentation of relevant information is often not made available to the wider public through the home domains municipalities or provincial government.
Opening slides of a smart strategy presentation hosted by the City of Cape Town alongside SAP technologies.
Without going into too much further gratuitous and quite mundane detail, the approaches with which I tested and experimented therefore did not seem a suitable fit to some of the contextual “textures” highlighted in this post. The more structured digital enquiry and methodological approaches tested could not accommodate the South African digital context in which they would be deployed.
Considering the described snapshot moments of working in this South African digital context, I would argue that a form of web crawling and digital enquiry that accommodates manually investigating and navigating this varied topography (a digital wild west where there are no prescribed routines or maps to navigate the undulating terrain) is essential.
Surfacing the information on how smart city discourse lands and unfolds in South Africa is clearly not a cut-and-dry manner. As with researching and sufficiently navigating the uneven, chaotic, make-do nature of South African governance in motion, a fair amount of manual labour, ingenuity, and a useful dash of existing knowledge of the context is a prerequisite. A more adaptable, explorative digital approach is thus perhaps more well-suited than an automated, pre-programmed approach.
More programmatic approaches may produce results that flatten the unique topography, losing much of the nuance that adds essential insight into how smart city discourse lands, manifests and mutates over time in South Africa; in both the governmental digital space and the physically manifested, “IRL” space it represents.
An experiment of manually plotting of interconnecting relationships between smart city discourse-relevant actor-networks, as evidenced through City of Cape Town’s “digital strategy” presentations to external parties.
As a proof of concept experiment, this mapping notably demonstrates that manually crawling and processing can pick-up and highlight the subtleties and nuance in the South African governmental digital landscape; this is something not picked-up on with automated mappings by programs such as Issuecrawler.
With this post I am advocating for a digital method approach that stays true to the core intent outlined in my previous post:
How do narratives around smart cities land as a set of ideas in local governments, and how do they begin to mutate and grow across time?
As such, I am hesitant to adopt a methodological approach that does not attempt to retrofit a prescribed methodological intent, thereby creating needless complexities the investigation process and adding unnecessary ad hoc processes that are not informed directly by the context. Whether this context is “real” or digital is irrelevant, it serves as a context nonetheless, one from which the research should draw it’s approach and not vice versa.
It is an approach to working with digital topographies for web crawling and scraping that pays heeds to the nuances that constitute more micro-scale digital textualities; textualities that construct context.
As a final statement to draw these couple of methodological pieces to a close, I issue a statement intended to challenge the methodological assumptions we make with our work. Taking the brief and selected insights into the South African governmental digital landscape, and giving time to consider the nuanced understanding necessary to formulate a suitable methodological approach; what of the other research site for the broader project?
When considering performing a similar digital enquiry into India, would it be fair to impose the same methodological approach for the sake of parity and consistency? If one is to consider the variances in the digital topography of South Africa, and considering the immense differences already seen in India from a mere cursory inspection of its own governmental digital space, is it ethical to impose such a methodological parity? Would it perhaps be a better approach to take time to consider India’s governmental digital topography and to formulate an enquiry and approach that is suited to its own digital context?