Computers, Society, and Nature

Graeme Hugo here elaborates a wide-ranging argument for the relevance of population geography to the question of international migration among the increasingly migratory populations of East Asia.
I found the article speaks to the increasingly complex global system of flows, encompassing goods, capital, and ideas as well as humans. Amidst what I’d call a general weakening of the sovereignty of states and a concomitant increase in their interdependence, the world of humans and their stuff resembles one unified system more obviously than ever before. At the same time, this world is massively chaotic, and while it was at some point relatively simple to analyze European immigration to America as a function of birthrates, automation of rural farm labour, and the growth of the American economy, the cyclical, multi-directional flows of human beings in and out of Asia in the 1990s rightly (as Hugo demonstrates) demands a different approach to understand.

And what of GI Science and big data? Hugo doesn’t delve as deeply into the complex methods he outlines as I would have liked, but with the multiplying ways that human beings can leave a recognisable trace today, I would argue that it has become generally easier to track even undocumented migrants. As evidence I’d present the exhibition “Forensis,” presented at Berlin’s Haus der Kulturen der Welt in 2014, which documents a multidisciplinary evidence-gathering effort undertaken to prove that NATO warships intentionally ignored a sinking ship full of African migrants. The researchers used advanced statistical methods, remote sensing data, modelling and visualization techniques, as well as human rights law to successfully mount a case against NATO in international courts. As we hone our techniques for detecting human beings, questions of our responsibility for them are naturally raised. http://www.hkw.de/en/programm/projekte/2014/forensis/start_forensis.php

Coming off of the heels of our discussion last week’s seminar, I can’t help (for better or for worse) to read the articles about geo-complexity through the lens of uncertainty. In particular, I am reminded of when Professor Sieber challenged me to make an argument for why uncertainty could be good, and I proposed that some level of geographic uncertainty is likely to mitigate the worst effects of spatially occurring trends of discrimination (e.g.: red-lining, gerrymandering, etc.), while also accommodating a diversity of geographic experiences and ontologies. In his article “Asia on the Move: Research Challenges for Population Geography”, Graeme Hugo discusses geocomplexity as it pertains to conceptualizing and analyzing human migration in Asia. I wonder–somewhat contrary to conventional wisdom–if we are headed to a world of more geographic uncertainty, in spite of the emergence of big data and the discussion of a “major and focused multidisciplinary research effort” in order to circumvent the “huge gaps in our knowledge of the patterns, causes and consequences of international migration in Asia” (Hugo 95).

Hugo points out that census data is predicated on the assumption that “individuals and families have a single place and country of residence”, and therefore are increasingly difficult to use for studying migration patterns. As discussed in the paper, ease of travel has accommodated several migration patterns which involve living part-time both in the nation of origin and the nation of desitnation. Although Hugo presents several secondary sources for understanding migration trends, he notes nonetheless that understanding migration patterns is complicated by the increasing volume of migration, as well as the “increasing heterogeneity of the international labour flows and the people involved in them” (Hugo 103). It is that remark about the “heterogeneity” of labour flows that intrigues me.

If the motives behind labour migration are increasingly divergent, what implications does that have on studying migration patterns at all, even if we develop techniques to use secondary/alternative sources to mitigate the issue of geocomplexity? In my opinion, this will mean that certain assumptions held by human geographers will become invalid; in the case of migration and geocomplexity, this will mean that we cannot assume the migration was necessarily driven by economic necessity. Increasingly, rich, middle-class, and poor people are drawn to migrate for a variety of reasons, and even if we grasp exactly how many people move around, we will not be able to make assumptions as to why, or even the nature or duration of their migration.

To frame this another way, even if the quantity of data we are collecting is increasing, I believe the certainty, validity, and utility of it is often decreasing. In the same way we’ve discussed the limitations of making sweeping demographic assumptions about VGI (e.g.: people post information on social media selectively and aspirationally), so to are their limitations of capturing migration patterns in any region of the world. The reasons for migration are increasingly heterogenous and simply having numbers tells us nothing. In my opinion, this is bad news for Uber, Facebook, or any other company whose stock market value is intimately tied to the anticipated value of their amassed datasets. But it’s good news for anyone who’s worried about their privacy and ability to be profiled by their data footprint. It’s certainly contrary to the general thrust of this course, but I think our ability to be profiled based on data footprints is overstated.

~CRAZY15

Even from the first page, I received the impression that “Complexity theory in the study of space and place” by Manson and O’Sullivan (2006) was a well-written paper. It tells a story and proceeds at a smooth pace that is easy to read, while still providing substantial information on the topic. I did find the constant references to various philosophical theories, such as reductionism and holism, difficult to assimilate into my understanding of complexity as I do not have a background in such theories. I felt like I was receiving an introduction to philosophy and complexity at the same time – a bit overwhelming! However, it did make me realize that an understanding of basic philosophical theories would probably help my conceptualization of GIScience as a whole – which was not a connection I thought to make in this class. To give credit where it is due, the authors did help comprehension by providing short definitions or context for obscure words within the text.

When asking the three main questions of “(1) Does complexity theory operate at too general a level to enhance understanding? (2) What are the ontological and epistemological implications of complexity? And (3) What are the challenges in modeling complexity? (678)” the paper highlights the tension inherent in the field of complexity. One problem that seemed especially prominent was the conflict between understanding emergent behaviour and the desire to simplify models. Computational modelling was provided as both a solution to accommodating large amounts of heterogeneous variables while also being presented as an easy avenue towards simplification (683).

The authors also made some references to spatial scale that I found particularly intriguing – namely how emergence and scaling up from local to more global phenomena can conflict with modeling assumptions of uniform patterns over different scales. I am finding more and more that all of our individual research topics are converging on each other. Complexity relates to spatial scale, which relates to ontologies, which relates to uncertainty, and so forth. I have not yet fully decided what that means for the broader context of understanding GIScience in my own head but I think it is important to acknowledge the increasingly common ground. I feel as if, through this class, I am step by step building my own conceptual network model of GIScience. It is not a linear path by any means – rather circular and backtracking in fact – but slowly, slowly, slowly the connections form.

-Vdev

Excuse me for my invocation of a bit of prose, but this was the first thing to spring to mind upon completing the article by Jones et al. (2008):

What’s in a name? that which we call a rose

By any other name would smell as sweet;

So Romeo would, were he not Romeo call’d,

Retain that dear perfection which he owes

Without that title. Romeo, doff thy name;

And for that name, which is no part of thee,

Take all myself. (2.2.47-53)

– Juliet, Romeo and Juliet

Personally, I usually can’t stand the insipid characters in the aforementioned play but in this case they do provide interesting context. While Juliet is happy to ignore all the meaning in a name, I would argue that Jones et al. do the opposite – in fact, they assume that all names have meaning enough that even vague geographic descriptors should be subject to quantitative analysis. I do wonder how big data can help analysis of vague spaces because of the sheer quantity of data that could allow for better modelling.

After our discussion in class about indigenous use of GPS and GIS, I do also want to know how modelling of vague places could be specific to an ontology that prioritizes precision. Does everything need to be quantified? Should it be quantified and where exactly does the role of ambiguity play into more cultural understanding of places?

Furthermore, I really liked the description of how a place can be described by what it is not, for example an area can be clearly defined by determining all boundaries around it. A fun fact, the Student Society of McGill University was actually called the “Students’ Society of the Educational Institute Roughly Bounded by Peel, Penfield, University, Sherbrooke and Mac Campus” as a protest against not being able to use the word ‘McGill’ in Student Clubs. Overall, names are incredibly important and can be described in many ways and methods of quantifying vague names could give rise to new understanding of how space is conceptualized.

-Vdev

In “Modelling vague places,” Jones, et al. introduce a novel method of natural language processing for vague toponymic data. They use open-source Named-Entity Recognition methods to extract associative place-names from the results of Google searches of vague toponymic terms such as “the Cotswolds,” an area straddling 6 different counties in Southern England. Then, using a gazetteer, they assign coordinates to the data extracted to transform the text into geolocated points. These are interpolated using density estimation techniques to draw the boundaries of vaguely-defined regions.

The process is representative of the general move toward big-data research: in the past, researchers on the topic would conduct interviews with a necessarily limited number of human beings who would sketch out their notions of boundaries or centres of vague areas. Meanwhile, GIS systems employ administrative definitions which are clearly not always suited to the needs of, say, a google-maps end-user who wants to know the boundaries of a neighbourhood such as Mile End, which has no official representation on a map or spatial data layer. Ask 10 different Montrealers where the southern boundary of the neighbourhood lies, and you will probably get several different answers. If an ontologically precise boundary definition were the goal, we might prefer the huge n-value of this sort of textual analysis to the anecdotal reports of several different people.

While the researchers employ a gazetteer to assign geographic coordinates to place-names, we can imagine that geolocative metadata extracted from Facebook posts or tweets could offer a potential alternative, especially when dealing with small, densely-populated areas of cities rather than large regions like the Cotswolds or Scottish Highlands.

I imagine that big-data approaches offer a lot to the development of natural language processing–the ability of machines to process language as humans do. In some areas of NLP, such as named-entity recognition, machines can almost match humans’ ability to determine which words signify a person, an organization, or a place. As computers become better at thinking like us, they may begin to teach us the “truest” meaning of our own concepts.

-grandblvd

After reading the paper on approaches to uncertainty, it was interesting to see a case study of how these concepts are put into practice. In Approaches to Uncertainty, the authors outline the nature of uncertainty in spatial data. The authors outline two strains of uncertainty, one strain where the object is well defined and therefore the errors are probabalistic in nature. The other strain of uncertainty is when the object is poorly defined, which results in more vague and ambiguous forms of uncertainty. In Modelling vague places the authors describe a method of density modelling as an effective method of representing the uncertainty of a place name extent.

In Jones’ article, they discuss the difficulty of storing spatial information for vague place names like the Rockies, or a downtown, that are not strictly defined. The authors mention that when trying to determine how subjects conceptualize vague places, interviews are a powerful tool. They then go on to conclude that automatic web-harvesting is a better way to go, because it is clearly more time efficient. However there is still room for interviews in smaller scale studies.

For example, I found this article to be a good addition to a discussion on qualitative GIS that I had in GEOG 494. In the presented reading, a researcher had collected information through interviews of how safe a Muslim woman felt going about her daily activities through space post-911. Through her interviews, the researcher found that her subject’s definition of friendly space had shrunk in the post-911 society. I just bring up this example to show that not all uncertainty due to vague definitions of space in a GIS can be modelled using web-based knowledge or automated processes.

-anontarian

I found the chapter “Approaches to Uncertainty” to be an interesting read, although it is definitely one coming from an empirical, quantitative perspective. In particular, the discussion of ambiguity was interesting and somewhat confusing to me. I think that even the existence of discord depends on the user, the individual defining the object. In a territorial dispute, one individual may not even recognize that a dispute exists, while another might argue over it. Something that I found difficult about the author’s discussion of discord was that in the flow chart (figure 3.1), “expert opinion” follows from discord. This seems troublesome to me, and does not seem to fit with the example the author uses for discord. In a land dispute, where two groups have laid claim to an area and have deep roots there, it would not be appropriate to have an expert’s opinion. For one thing, who would be the expert? There are power dynamics inherent in who resolves spatial uncertainty, and in doing so, legitimizes one thing or another.

The article also made me reflect on our discussion about indigenous epistemologies. Rundstrom (1995) describes how indigenous people exhibit a “trust in ambiguity” and embrace the nuances of geographic spaces and living beings. In the article, ambiguity is defined as confusion over how a phenomenon should be classified because of differing perceptions of it. I think indigenous people as Rundstrom understands them would take issue with “how” a phenomenon should be classified, and argue if it should even be classified at all. Can GIScience embrace ambiguity in some ways? There is certainly a need for a way to incorporate more ambiguity into GIS if we are to try to represent indigenous geographies.

-denasaur

(As a side note if anyone is interested: I thought that the article at the following link brings up some interesting questions about spatial uncertainty – it incorporates many of the definitions this article does, as well as some discussion of indigenous conceptions of space. The figure 1 diagram is a good visual. http://pubs.iied.org/pdfs/G02958.pdf.)

Wang’s article is on cyberGIS; software that operates on parallel and distributed cloud-computing rather than the typical single-computer sequential GIS. This software, particularly the CyberGIS Gateway, reminded me of our class discussion on how GIS is taught in a research-oriented university. Initially I was frustrated that research-oriented schools are apprehensive to teach a step by step class on using ArcGIS. However from a practical perspective it doesn’t make sense to teach one software when there exists so many open-source softwares that catch-up quicker to the demands of researchers and businesses. It appears that our computing capability, especially through cloud computing, is increasing so rapidly, as are our datasets, that CyberGIS will become the future.

As we move towards CyberGIS the hope is that the traditional cost and skill barriers to GIS will fall, opening up the toolset to a wider range of disciplines. Then should the challenge of CyberGIS be to make the toolsets easier for this wide range of people to use? The development of easy tools for hydrology and emergency management allows decision makers to access powerful networks of computers to manage complex data. In this sense I think CyberGIS can empower smaller organizations to make good decisions, rather than corporations or governments. Often big-data applications are criticized as being the tools of large corporations like Google possessing expensive infrastructure. CyberGIS allows smaller organizations to compete with large firms and possibly break down the hegemony of these massive companies by performing complex analytics without the expensive infrastructure.

-AnOntarian

Yang’s paper is an exhaustive review of the advancement of the Cyberinfrastructure that has grown since individuals desired to define such a term. Yang et al discuss both the utilized and untapped potential of the interconnectedness of the world in the 21^st century, first from a general perspective, and then a Spatial/Geographic perspective.

As the authors discuss the existence of this network, I found that the desire to define the term came after the inception of the network and from the desire for connectivity among the vast amount of info available. It seems that the main theme of this paper is intelligent integration and cooperation between various computing platforms and scientific agencies, such as NASA, NEPTUNE, and ESRI.

I found Yang’s review of the GCI’s untapped Environmental and Geographic potential to be the most accessible and obvious component. Ideals such as heterogeneous integration between various institutions and data collectors multiply the analytical possibilities of scientific research. From a Geographic perspective, I believe a healthy GCI is the next logical step in the evolution of GIScience, following the inception of the Geographic Information System. With the introduction of big data and open source information, individual users and consumers can become more involved in a field that would be otherwise inaccessible if not for the existence of GCI’s to simplify data intensive endeavors, such as those discussed in section 5.9, Education.

Smitty_1

The title of the article by Yang et al. (2010), “Geospatial Cyberinfrastructure: Past, present and future” should have given me a clue as to the extensive scope of the paper. By trying to cover almost every single aspect of GCIs, the authors provided an impressive review. Yet it was somewhat overwhelming as an introduction to the subject. I see the value in this paper as a reference text for more knowledgeable users. However, I would have liked to see more concrete, in-depth explanations of GCIs. I think my understanding of a GCI would also have been aided by an in-depth description of an unrelated CI and how exactly it was different from a GCI. Essentially, more tangible references would have helped my comprehension. The authors themselves link their work to GIScience by stating that this is a review of recent developments and that “similar to how GIS transformed the procedures for geospatial sciences, GCI provides significant improvements to how the sciences that need geospatial information will advance (265).”

While the article was very clear about the direction that GCI advancement should go in, the authors skimmed over barriers that might impede progress towards those end goals. The desire in particular for “a semantic (ontology) based framework that is sensitive to the scale, richness, character, and heterogeneity within and across disciplines (272)” is almost a chimera. I would argue that the ‘grand challenges’ briefly identified should be expanded into full papers themselves. How to integrate cyber infrastructures across disciplines and shift them to be human-centered paradigms are challenges that, once solved, could provide substantial improvements to the field. Geospatial cyberinfrastructure development seems to be at a crucial turning point. If all contributors could individually maintain as thoughtful a vision of the GCI framework for the future as Yang et al. while resolving current discrepancies then these far-reaching goals might become attainable.

-Vdev

Throughout the course, we have discussed various ways in which GIS can be manipulated for unethical ends. For instance, we have asked: to what extent does online advertising which discriminates based on assumed demographic characteristics exclude marginalized populations?; to what extent can business practices (such as Uber’s “surge pricing”) contingent on GIS data be considered appropriate?; and how does military involvement in the development of GIS operations implicate the field? In their article “Geospatial Cyberinfrastructure: Past, present and Future”, various goals at making CGI more inclusive, democratic, and multi-disciplinary are expounded upon. For instance, we are promised that CGI will aid “to advance citizen-based sciences to reflect the fact that cyberspace is open to the public and citizen participation will be essential” (264) and provide a standardized way for a multitude of actors including “government agencies, non-government organizations, industries, academia, and the public” (264) to manipulate geospatial data. Yang et al argue persuasively that the complexity and interdisciplinary scope of contemporary problems such as developing “strategies to reduce energy consumption and stabilize atmospheric emissions so that global temperature will not increase… [and choosing] a housing site that minimizes the risks of forest fire, flooding and other… hazards” (267) demands a coordinated approach and that CGI–with its enabling technologies such as web computing, open-source software, and interoperable platforms–is able to provide a coordinated platform for this problem-solving. But in this push to make GIS simultaneously more democratic and legible to a variety of actors, how will GIS remain an ethical science? Already in the “closed” world of GIS where meaningful operations require access to knowledge and resources, energy companies have assembled legions of capable GIS technicians to explore for extractable resources, and companies have established marketing departments engaged in ethically dubious GIS practices. So what does the world look like once the barriers of cost and knowledge to GIS use are removed or, at least, decreased? Do the ‘good guys’ win their case more often because they now have access to a multitude of data once available only behind a walled fortress of GIS elites? Or does the ease of access allow the data to go completely unmonitored its use? In other words, if we continue to hold that GIS is a science, how can the field maintain a ‘soul’–its own Hippocratic Oath, if you will–and maintain a reasonable set of ethics best practices? And how does it remain a science when its increasing scope and level of interoperability will have many academics and non-academics using it primarily as a tool?

-CRAZY15

In his 2015 paper, Shaowen Wang outlines the current state go CyberGIS as a growing ‘interdisciplinary field’ that hopes to enable widespread cooperation on geospatial research by creating a framework which integrates all sorts of data processing techniques from a number of ‘research domains’ and allows for real-time high volume, multi-user work by taking advantage of modern day networking technologies such as cloud computing, multi-core processing on an unprecedented scale and remote work.

At first, CyberGIS felt like a catch-all umbrella term with little purpose to me, a buzzword that packaged old concepts in a new way. Following Wang’s article however, I am convinced of the relevance of CyberGIS and the exciting possibilities it offers, particularly with regard to scaleability and the democratization of computer processing power.  Our in-class discussion of the paradigms which narrow our understanding of GIScience (e.g. attention to ‘maps’ over all other forms of data presentation) informed my reflection on Wang’s review of the existing status quo in research: ‘sequential computing’, ‘monolithic architecture’ and other concepts which we take for granted when we engage with online work.

CyberGIS has the potential to become an uprecedented force for radical change in academic research. The notion that a researcher defaults to working alone from a modest work station, carefully guarding the fruit of their research and selectively collaborating through individualized in-person or online communication could soon be overhauled. The barriers to research inherent in the limited access to powerful computers could potentially be broken down by a combination of cloud computing, easier collaboration and delegation of tasks, and real-time remote access to more computers/better facilities. This has enormous implications for the democratization of GIScience which go hand in hand with the tenets of the open source movement. Together, CyberGIS and open-source approaches to software development could certainly change the nature of resource accessibility in academia, providing opportunities for better collaboration and less elitism in GIScience. In my opinion, GIScience continues to suffer from high barriers to entry on both a physical and intellectual front, and it is high time for a fresh approach. CyberGIS might just be it.

-KH