geog595
geog595 copied to clipboard
jakobzhao
04-data
Kwan's paper (2012) discusses the uncertain geographic context problem (UGCoP). They started by talking about interlinkages and gaps between health studies and geography where they pointed out that public health personnel may know very little about spatial distribution and boundary, leading to some research bias. Most of the research is based on area-based contextual attributes, such as the census tract, which is usually the easiest or the only viable way to conduct the study. However, the disadvantage of this area unit is that it may not well-represented of the actual space's influence on an individual. People interact with larger geographical areas and also social connections are often happening beyond the geographical boundaries. Therefore, there is a mismatch of the boundary or scale of the space. I really like the conclusion that made that "humans are active agents who construct their own geographic contexts and tie together different spatial scales through their daily activities, movements, and social interactions." It reminds me of my own research, wherein a traditionally agricultural community in Chengdu Plain China. People go to the nearest rural market not only based on the physical distance (i.e. they don't usually go to the nearest market). In contrast, they choose to go to the rural market where they are most familiar according to their social connections (i.e. their neighbors also go to that market), or it is because the market is located in the same administrative village so that apart from doing shopping, they can also go to the village community, the village hospital or pharmacy and do some own businesses. This phenomenon can be explained by one of Kwan’s arguments that “the interconnections among individuals and places are vastly complex and vibrantly dynamic, and they should be conceptualized and examined as such.”(p.966) MacEachren et al's paper (2005) talks about uncertainties of information. It associates strongly with the first paper that many information uncertainties are generated based on human individuals. The human individual cannot be treated as a complete "reasonable" or "objective" agent; instead, they are "subjective" or "emotional" agents who do not behave like machines and therefore brings along with a lot of uncertainties. How to delineate such uncertainties is an ongoing topic in the science community. The third paper by Zhang and Zhao (2020) discusses the transformation of today's data landscape. They first identified the different terms between "spoofing" and "fake or false". Followed by that, they discuss the real-world example of "standing rock" which is a campaign that reflects the geographical location spoofing on social media. Some methods they use including Python crawler, word cloud, and Qgis. They stated that instead of information passing from top to down, more “grassroots” information could trigger a movement from bottom to up and bring the data landscape into a new “post-truth era”. The authors cite Foucault's suggestion on “democratizing the epidemiology of truth”. Finally, they argue that "our job as intellectuals is probably not to build a new regime of truth but first to recognize and understand new kinds of truth politics that are already emerging from the bottom up as in this case of a spoofed check-in protest on social media".
04/26 Steven Bao
This week's readings really make me realize the academia's efforts to find truth in both its objective and subjective forms.
MacEachren et al. (2005) focus primarily on (geospatial) uncertainties in their traditional definition, which is mainly objective. Data is considered certain when it could be measured and corresponds to the physical object that it represents or describes; otherwise, we have uncertainties. In my opinion, this paper is "data-centric" because it presents frameworks for understanding and visualizing uncertainties within data so that the human users from different disciplines could well understand the "uncertain data" as well as the possible influences it may bring.
In 2012, Kwan articulated the "The Uncertain Geographic Context Problem," a fundamental and inherent problem within the common methodological approaches in scientific research. The problem is caused by the assumptions made by researchers, either intentionally or unintentionally, about how, or to what extent, spatial contextual variables interact with the outcome variables. Such uncertainties are hard to visualize using traditional visualization techniques, yet, undoubtedly, they require improvements on the existing scientific research procedures for making right scientific inferences that reflect the truth.
As data continues booming in the "Big Data Era," Zhang et al. (2020) re-conceptualize "truth" based on contexts and advocate for an epistemological shift when working with the fast-emerging data. According to Zhang et al., user-generated "fake" data can also be meaningful and reveal "truth." The uncertainties within the "truthness of data" could only be understood if we include and consider the contexts of the data, even though the contexts may also bring subjectiveness. This is particularly important in today's technology environment, where social media information and machine learning are increasingly adopted in social science research. Understanding the contexts of user-generated data could help identify the underlying messages within the data. Either as people who use machine learning to generate data or as ones who collect data from the products of machine learning, while the "garbage-in-garbage-out" principle remains essential in machine learning, it is more important for us to consider carefully what is "garbage" and what is not.
References: Kwan M-P (2012) The Uncertain Geographic Context Problem. Annals of the Association of American Geographers 102(5): 958--968. DOI: 10.1080/00045608.2012.687349. MacEachren AM, Robinson A, Hopper S, et al. (2005) Visualizing geospatial information uncertainty: What we know and what we need to know. Cartography and Geographic Information Science 32(3): 139--160. DOI: 10.1559/1523040054738936. Zhang S, Zhao B, Tian Y, et al. (2020) Stand with #StandingRock: Envisioning an Epistemological Shift in Understanding Geospatial Big Data in the "Post-truth" Era. Annals of the American Association of Geographers: 1–21. DOI: 10.1080/24694452.2020.1782166.
Jenny Lee
04/26
What counts as credible geographical data? What are the boundaries to geographical data? Both Zhang et al. (2021) and Kwan (2012) approach these questions through a human centered contextual approach to the construction of geographical data. Kwan (2012) however seems to ground the analysis in a more post-positivist approach than Zhang et al. (2021). Kwan’s approach reflects the spirit of correspondence theory introduced in Zhang et al. (2021). Her article is concerned with resolving the “uncertain geographic context problem (UGCoP), which she defines as problems that arise from the mismatch between the geographical context measured by the research and the “true causally relevant geographic context” (Kwan, 2012, p. 959). Kwan (2012) aptly explains that peoples’ movements and interaction with others, which are important contextual variables that affect peoples’ health, do not have clearly demarcated boundaries. Therefore, research that simply rely on stagnant residential data to study geography’s effect on health overlook the complexity of contextual factors that impact peoples’ wellbeing. As a remedy, Kwan suggests that researchers construct a theoretical model that can incorporate the “spatial and temporal characteristics” that closely resembles that of the “true geographic context” (p. 963). In this aspect, Kwan’s research still relies on the assumption that true geographic context is out there. The contextual approach of Zhang et al. (2021) challenges the epistemological assumptions about truth as a reflection of scientifically observable reality and the binary understanding of data as true or false. Through a case study of the #Standingrock movement, the authors introduce location spoofing, “a way to claim one’s location inconsistent with one’s physical location” (p. 2) and explore how this bottom-up construction of geographical data illuminates the importance of incorporating peoples’ intention into contextual cues.
Particularly noteworthy is their incorporation of the Actor Network Theory into their contextual approach: The authors outlined and examined how the “socio-technological dynamics between the involved actors” (p. 2) create the context of the #standingrock movement. Their article also reminded me of Sophia Baik’s (2020) conceptualization of the #standingrock movement on Facebook as a “geotagging counterpublic” made up of users who strategically utilize their geolocation data for political organization. The geotagging counterpublic could be said to counter the dominant discourse surrounding “verifiable” understanding of the movement advanced by institutions and traditional news media as mentioned in Zhang et al. (2021).
I found Zhang et al. (2021) article especially helpful in envisioning different ways of studying digital space and look forward to applying their approach to my own research. While Zhang et al. (2021) explored the bottom-up construction of truth, I am particularly interested in the bottom-up construction of surveillance across our networked media landscape. This makes me think about whether the texture and surveillance through spatial expansion on digital space is fundamentally different from physical monitoring, and how that would change the epistemological assumptions about power in digital spaces.
References
Baik, J. (2020). The geotagging counterpublic: The case of Facebook remote check-ins to Standing Rock. International Journal of Communication, 14, 2057–2077. Kwan, M.-P. (2012). The Uncertain Geographic Context Problem. Annals of the Association of American Geographers, 102(5), 958–968. https://doi.org/10.1080/00045608.2012.687349 Zhang, S., Zhao, B., Tian, Y., & Chen, S. (2021). Stand with #StandingRock: Envisioning an Epistemological Shift in Understanding Geospatial Big Data in the “Post-truth” Era. Annals of the American Association of Geographers, 111(4), 1025–1045. https://doi.org/10.1080/24694452.2020.1782166
The papers this week certainly draw us into questions of epistemology, questioning how we know what we know and how we, as researchers, grapple with truth and uncertainty. Kwan’s mention of the modifiable areal unit problem immediately called my current project to mind (an attempt to overlay an red-lining map from the 1930s with current Census tract data from that same area), but this paper was an exploration of a different problem that geographers face: the problem of uncertain geographic context problem (UGCoP). More simply put, this is that people move! The spatial and temporal variability of people’s activities makes causality especially tricky to establish in relationship to a defined geographic area. Kwan’s acknowledgement of that the “interconnections among individuals and places are vastly complex and vibrantly dynamic interconnections among individuals and places are vastly complex and vibrantly dynamic (p 966) is useful in thinking through research design of any type since there is detailed discussion of (un)certainty, logical fallacies, and overreaching conclusions, all of which can plague researchers in any discipline. While Kwan suggests the use of GPs to counter UGCoP, another way that I think this methodological conundrum might be addressed is for a researcher to employ a mixed-methods approach that grounds the research in context. Zhang et al remind us that the “old guard” of truth and information (they name scientists and politicians) is no longer as powerful, due to the decentralized dissemination of data, but take this further to explore the process of truth-making itself through the examination of what digitally transpired during the Standing Rock protest. They caution the reader that simply dismissing data as “false or fake” misses an important factor in understanding the broader context: the human intentions behind data. This paper is particularly interesting because of their utilization of anthropological methodology in the space of what could be understood as “the technical” and also in that this interpretation bridges GIScience with philosophical perspectives of truth as stated here: “Two major existing GIScience strategies, mechanical and statistical, follow the same epistemology rooted in the correspondence theory of truth, despite their strategic differences, in which one insists on a single true value, whereas the other estimates a range that contains true measurements.” I appreciated their interpretation of data as both containing information and intention and, with last week’s Winner reading still front of mind, am thinking about data in the same way as artifacts: pluralistically.
The enhancement of digital technologies has brought about the new regime of truth and data featuring big data and the involvement of human intentions. This entails uncontrolled data quality and uncertainties. The emergent characteristic has occurred in the "post-truth" era. Post-truth politics is a political culture featuring appeals to emotion disconnected from the details of policy and ignoring factual rebuttals (Wikipedia, 2021). Deflationary theories such as performative theory and pragmatic theory feature human intention for the new regime of truth. This intention entails the fitness-to-purpose presenting data actor's intentions as opposed to fitness-for-use, which is characterized by accurate measurement of the correspondence theory. Furthermore, post-truth politics are now challenging to political and technological authorities due to the enhancement of digital technologies such as social media empowering users to be significant actors while producing big data. These bottom-up initiatives are powerful enough to question epistemological frameworks for the new regime of truth or data. For example, location spoofing presents a challenge to the authorities, who used to shape public opinion and political outcomes, with the bottom-up data actor's intentions. This can be another political regime of truth requiring reconfiguration of the previous one (Zhang et al., 2021).
MacEachren et al. (2005) suggested conceptualizing uncertainties for visualization of information uncertainties. The authors distinguish between error and uncertainty according to whether inaccuracy is known. First, they introduced the characteristics of data quality - lineage, positional accuracy, attribute accuracy, logical consistency, and completeness - and uncertainty - data/ value, space, time, consistency, and completeness. Then, they suggested data uncertainty characteristics - accuracy/ error, precision, completeness, consistency, lineage, currency, credibility, subjectivity, and interrelatedness followed by challenges.
Data collection also involves uncertainty. Uncertain geographic context problem (UGCoP) - how contextual units are geographically delineated leading to deviation from the true geographic context for the effects of area-based attributes on individual behaviors. This makes the examination of the effect of contextual influences on individuals such as health confused. More appropriate contextual units based on people's actual activity spaces are necessary as the contextual unit varies over space and time depending on the population groups as opposed to the conventional units such as residential neighborhoods, social contexts, administrative areas, etc. Using GPS may delineate activity space by measuring the spatiotemporal variation of the contextual unit on individuals to address UGCoP. This new mobilities paradigm features movement, mobility, and space-time capturing daily activities, movements, ad social interactions of individuals (Kwan, 2012).
As a student of Sociology, there is a principle that seems relevant to the issues we are discussing in this course: “Innovation is the mother of necessity” (Nolan and Lenski, 2015). The earliest humans began innovating (making tools/weapons) to satisfy their basic needs. As humans continued to innovate, the benefit of the previous advancement gave rise to new issues to deal with. For example, the plow revolutionized horticulture and agriculture making it possible to produce more food and sustain a larger population. At face value, it is pretty straightforward, but what did these revolutionary innovations breed? The necessity for better and more stable housing, a more complex social stratification system, a warrior class to protect the now sedentary settlements and food surplus. A similar pattern of innovation requiring further advancement to address both the unintended and intended results continues to this day, as evidenced in our readings this week.
The advent of the digital space opened up infinite possibilities, and the more people gained access to that space, the more opportunities there were for even more innovation. But, this is not a straightforward linear pattern, there are unintended consequences (they don’t have to be good or bad) that complicate the part of the world made accessible by the innovation of the digital space. Increased access to social media and digital capabilities also increases the interaction between digital tools and natural human proclivities. Given the complexity of humans, it makes sense that “In online Environments, user-generated data cannot be taken at face value” 17 (Zhao & Zhang, 2020).
The issue of human complexity and human engagement with the digital came up for me again while I was reading MacEachren (2005). What seemed to be missing from this piece was the consideration of human psychology. Psychologists and neurologists have plenty to say about how the human brain processes information to make decisions and the mechanisms within that process. This consideration of human neurology seems integral to making any progress in helpfully communicating your data and its uncertainty, especially when you are attempting to give such a specific kind of information to your reader that is intimately and uniquely connected to your area. Again, this is an example of an innovation raising more questions, but one thing I am confident saying for certain is that the technology will only be clarified through its relationship to humans, and we are hardly ever cut and dry.
• MacEachren, A.M., Robinson, A., Hopper, S., Gardner, S., Murray, R., Gahegan, M. and Hetzler, E., 2005. Visualizing geospatial information uncertainty: What we know and what we need to know. Cartography and Geographic Information Science, 32(3), pp.139-160.
• Nolan, P., Lenski, G. 2015. Human Societies: An Introduction to Macrosociology. Oxford University Press, 12th edition pp. 52 -86
• Zhang, S., Zhao, B, 2020. Stand with #StandingRock: Envisioning an epistemological shift in understanding geospatial big data in the “post-truth” era
humans are active agents who construct their own geographic contexts and tie together different spatial scales through their daily activities, movements, and social interactions."
there is no "being" of beings. Each individual should be learned as a human-in-the-world.
This week’s reading material shares a common train of thought about determining truth and uncertainty. These questions are fundamental yet complex to find a middle ground because of their subjective nature. Kwan (2012) presented to us an uncertain geographic content problem that describes the inconsistency between the measured geographical context and the true causally relevant geographic context. The issue of interest originates from people building their geographic contexts through daily activities, movements, and social interactions in space-time. These intrinsic dynamics render the analysis based only on static residential neighborhood data secondary since it leaves out the complexity of the spatiotemporal patterns of relevant contextual influence. Kwan’s argument that “the interconnections among individuals and places are vastly complex and vibrantly dynamic” reminds me of the previous think pieces where the interaction among people and space could add more meaning to “space.” How human and space interacts is evolving and carries its meaning that academia should not overlook.
Kwan’s work also compared the measured geographical content to the true context, and it has been a heated discussion in the other two articles about truth and uncertainty. Zhang’s work called for an update to epistemological assumptions about truth. I was touched by the sentence “simplistic labeling of data as false or fake can be misleading because it misinterprets or simply discards the meaning of human intentions behind data” in Zhang (2021). The quote was even more powerful when we considered the #StandingRock movement. People are intentionally introducing geographical information that does not reflect their exact locations at that moment for a moral purpose. That ethical purpose (and the ethical framework people choose to adhere to) is subjective indeed but with significant context influence. Also, consider a hypothetical case where a website uses bots to boost its content relevance to a keyword such that the website will be ranked higher and displayed “at the front” when someone searches that keyword. Is this action ethical if the content is some malicious ads or an emergency post of Amber’s alert? On the surface, both situations look like an abnormality in data. But can we treat the data as false/uncertain, and who can be the judge of making that decision? Zhang et al.’s effort to bring in human intentions is much appreciated.
This week’s readings two about uncertainty and one about location spoofing have a lot of connections with my own studies – “fake” phenomena in Geography. The first paper talks about the definition of geospatial uncertainty and the visualization of uncertainty. In Zhao’s paper about the difference between uncertainty, mistake, and spoofing (Zhao, 2016), uncertainty is to some extent the inherent vagueness of measurement, but in MacEachren et al.’s paper (2005), uncertainty seems a bigger idea: “When inaccuracy is known objectively, it can be expressed as error; when it is not known, the term uncertainty applies”, which not only contains measurement inaccuracy but also includes ideas like veracity and other unknown things. The broad concept of uncertainty is also reflected in Gwan’s paper (2012). Maybe I should have a section in my dissertation to discuss uncertainty, in case of ambiguity in use. MacEachren et al.’s paper also remind me of a recent read paper about truth and accountability in geospatial visualization (Wallace and Heuvel, 2005), in which the authors argued to include uncertainty information in the metadata of maps to reduce misunderstanding of maps and enable trustworthiness. I think directly visualizing uncertainty as described by MacEachren et al. (2005) can also be considered as a way to embed uncertainty on a map.
When it comes to Kwan’s paper, I’m to some extent shocked since one of my published papers about health issues and several past conducted projects just fall into the trap described as uncertain geographic context problem (UGCoP) by Kwan (2012), and I never realize it !!! For UGCoP, I also understand it as another form of improper proxy variables often referred to in econometrics. In the broader form of improper proxy variables, UGCoP remember me of an example of a rental house price study (Harten et al., 2020): the same horse price model using online big data and survey data generate a significant difference in the result: the online house location is not a good proxy of the reallocation and contextual environment of the house, since it contains the communication between householder and renter and always distort the location of a house.
Zhang et al.’s paper (2021) discussed location spoofing in #StandingRock using actor-network (ANT) (in the #StandingRock movements, thousand of users spoofed their twitter location to be in North Dakota supporting the local protests against a crude oil pipeline - here location spoofing is understood as a kind of democracy and resistance). This newest paper of location spoofing studies directly and sharply criticized the simplistic labeling of data as false or fake, understood truth as a social construction embedded with human intention, argued suspending moral judgment and turning to alternative epistemology of geospatial big data such as fitness-for-use assessment. I’m happy I finally understood this paper by tracing the development of studies related to location spoofing. In general, this paper serves as one most important references for my own research.
I’ve heard about the uncertain geographic context problem (UGCoP) for a long time but never dived into it until reading Kwan’s piece (2012) this week. It first introduced the concept of UGCoP and explained that it is different from the modifiable areal unit problem (MAUP). While UGCoP arises because of the spatial uncertainty in the actual areas and the temporal uncertainty in the timing and duration, MAUP arises due to the selection of spatial scale and spatial zonation (The scale problem means the size of spatial units chosen has an influence on substantive results. The zonation may not reflect the real lines of division, and thus can influence the variance and mean outcome of interest. MAUP is a source of statistical bias). At the first glance, I just thought MAUP is a part of UGCoP in the sense that MAUP has the same spatial problem, but the author insisted that the two are different kinds of problems. As I read on and think deeper, spatial uncertainty in the UGCoP means that the selected spatial unit can be under change (e.g., residential neighborhood, workplace) all the time and some social contexts or social interactions (e.g., families, friends, peers) are not geographically defined and thus hard to measure in space. This epistemological shift is very inspiring, and in particular, it calls for a turn to the focus ‘from location to movement, from place to mobility, and from space to space-time’. In my research, I also have a need to use a GPS device to collect spatiotemporal trajectory data and use a portable camera to capture the videos or photos of changing geographical environments. I think I also follow a “new mobilities paradigm”, trying to cope with UGCoP.
I also appreciate the piece written by Zhang, Zhao, Tian, and Chen (2020). It is interesting to read about different philosophical and social perspectives on truth and reflect on existing GIScience approaches to truth and human intentions in truth. It is inspiring to think in the way in which rather than simply considering the labeling of data as false or right, it would be better to describe ‘an intentionally generated inconsistency between claimed and physical locations’ as location spoofing. This means that it would be too absolute to view data as false or right and that human intentions behind data, which can be positive, need to take into account by researchers. This point is well illustrated by the #StandingRock Movement explained in the paper.
(MacEachren et al.,2005) mentioned lay users are more likely to rely on instinct rather than statistical evidence when dealing with uncertainties. Their argument reminds me of my research experience in Japan, where my advisor and I helped airline companies develop mitigation strategies for volcanic ash hazards. The airlines are not interested in probabilities, and they tend to suppress uncertainties, which is coined as "precautionary-based" approaches. Visualization of uncertainties was not the focus of my study then, even though we tried to use greyscale to represent uncertainties. The focus was crisis respond strategies, and we found airlines are interested in risk scenarios, like stress tests, and they are especially interested in "worst-case" scenario, though we admitted we are not confident in how bad the "worst-case" scenario could be, the "worst-case" scenario we presented was based on historical observation of the volcano activities and wind fields. They just want unambiguous information.
(Kwan, 2012) addressed the uncertain geographic context problem (UGCoP). I encountered this problem in a study project from a course in last quarter. My topic was casual relationship between public transport and carbon emission. As mentioned in the article, I used census tracts--operationalized as static administrative areas--as contextual units. However, when I did sensitivity analysis, I found my results vary with delineation of geographic units. Kwan mentioned UGCoP can be addressed by using individual activity space to approximate the true geographic context. This solution resonates with our reading materials in week 2, the quantity and quality of geospatial data has improved and increased dramatically in recent years, new framework and technologies are required to source, store, and utilize geospatial data.
(Zhang et al., 2020) questioned what is the definition of truth in light of the widespread use of location spoofing, a way to claim one's location inconsistent with one's physical location. In my opinion, it depends on how we understand the location spoofing. If we understand it as their physical location, then apparently the location is wrong, because they are not physically at the location they tag on their posts. If we understand the location spoofing as a context, nothing different from words voicing their supports for the standing rock movement, then there is no right or wrong, it is just an express of their opinion. And I don't think these users are intentionally trying to manipulate others to believe they are at the spoofed location, and they are doing no harm. The word "post-truth" stems from recent political debates, politicians and media trump up fake news to shape public opinions. The location spoofing does not fit this definition. First, it is a bottom-up civil "uprise" against authority. “Post-truth” refers to top-down manipulation. Second, the location spoofing does not trump up fake information. As I have mentioned, if we understand location spoofing as a context expressing the support for a social debate topic, then it is not fake information. However, the bottom-up location spoofing movement shares one thing in common with "post-truth": objective facts are less influential than appeals to emotion and personal belief. It is not a problem exclusive to location spoofing, it is a problem to us all in the era of Internet, where everyone can have a voice, where grassroots can challenge elites, where people live in filter bubbles. Sadly speaking, appealing to emotion and personal belief is the easiest way to win popularity, and the public is becoming increasingly polarized.
This week’s reading material shares a common train of thought about determining truth and uncertainty. These questions are fundamental yet complex to find a middle ground because of their subjective nature. Kwan (2012) presented to us an uncertain geographic content problem that describes the inconsistency between the measured geographical context and the true causally relevant geographic context. The issue of interest originates from people building their geographic contexts through daily activities, movements, and social interactions in space-time. These intrinsic dynamics render the analysis based only on static residential neighborhood data secondary since it leaves out the complexity of the spatiotemporal patterns of relevant contextual influence. Kwan’s argument that “the interconnections among individuals and places are vastly complex and vibrantly dynamic” reminds me of the previous think pieces where the interaction among people and space could add more meaning to “space.” How human and space interacts is evolving and carries its meaning that academia should not overlook.
Kwan’s work also compared the measured geographical content to the true context, and it has been a heated discussion in the other two articles about truth and uncertainty. Zhang’s work called for an update to epistemological assumptions about truth. I was touched by the sentence “simplistic labeling of data as false or fake can be misleading because it misinterprets or simply discards the meaning of human intentions behind data” in Zhang (2021). The quote was even more powerful when we considered the #StandingRock movement. People are intentionally introducing geographical information that does not reflect their exact locations at that moment for a moral purpose. That ethical purpose (and the ethical framework people choose to adhere to) is subjective indeed but with significant context influence. Also, consider a hypothetical case where a website uses bots to boost its content relevance to a keyword such that the website will be ranked higher and displayed “at the front” when someone searches that keyword. Is this action ethical if the content is some malicious ads or an emergency post of Amber’s alert? On the surface, both situations look like an abnormality in data. But can we treat the data as false/uncertain, and who can be the judge of making that decision? Zhang et al.’s effort to bring in human intentions is much appreciated.
Grace, I am glad that you were appreciating the importance of intention in the context of determining truth/false.
Jennifer, I am glad that you feel both Kwan and Zhang et al's pieces of work are beneficial but from different perspectives.
Yifan, I am glad that you can link different nodes together from the three papers, and find their relationships upon uncertainties and inconsistency. Your think-piece reminds me of similar experiences that I always learn new things after a re-read of a classic paper.
(MacEachren et al.,2005) mentioned lay users are more likely to rely on instinct rather than statistical evidence when dealing with uncertainties. Their argument reminds me of my research experience in Japan, where my advisor and I helped airline companies develop mitigation strategies for volcanic ash hazards. The airlines are not interested in probabilities, and they tend to suppress uncertainties, which is coined as "precautionary-based" approaches. Visualization of uncertainties was not the focus of my study then, even though we tried to use greyscale to represent uncertainties. The focus was crisis respond strategies, and we found airlines are interested in risk scenarios, like stress tests, and they are especially interested in "worst-case" scenario, though we admitted we are not confident in how bad the "worst-case" scenario could be, the "worst-case" scenario we presented was based on historical observation of the volcano activities and wind fields. They just want unambiguous information.
(Kwan, 2012) addressed the uncertain geographic context problem (UGCoP). I encountered this problem in a study project from a course in last quarter. My topic was casual relationship between public transport and carbon emission. As mentioned in the article, I used census tracts--operationalized as static administrative areas--as contextual units. However, when I did sensitivity analysis, I found my results vary with delineation of geographic units. Kwan mentioned UGCoP can be addressed by using individual activity space to approximate the true geographic context. This solution resonates with our reading materials in week 2, the quantity and quality of geospatial data has improved and increased dramatically in recent years, new framework and technologies are required to source, store, and utilize geospatial data.
(Zhang et al., 2020) questioned what is the definition of truth in light of the widespread use of location spoofing, a way to claim one's location inconsistent with one's physical location. In my opinion, it depends on how we understand the location spoofing. If we understand it as their physical location, then apparently the location is wrong, because they are not physically at the location they tag on their posts. If we understand the location spoofing as a context, nothing different from words voicing their supports for the standing rock movement, then there is no right or wrong, it is just an express of their opinion. And I don't think these users are intentionally trying to manipulate others to believe they are at the spoofed location, and they are doing no harm. The word "post-truth" stems from recent political debates, politicians and media trump up fake news to shape public opinions. The location spoofing does not fit this definition. First, it is a bottom-up civil "uprise" against authority. “Post-truth” refers to top-down manipulation. Second, the location spoofing does not trump up fake information. As I have mentioned, if we understand location spoofing as a context expressing the support for a social debate topic, then it is not fake information. However, the bottom-up location spoofing movement shares one thing in common with "post-truth": objective facts are less influential than appeals to emotion and personal belief. It is not a problem exclusive to location spoofing, it is a problem to us all in the era of Internet, where everyone can have a voice, where grassroots can challenge elites, where people live in filter bubbles. Sadly speaking, appealing to emotion and personal belief is the easiest way to win popularity, and the public is becoming increasingly polarized.
Ziyang, thank you for your reflections, especially your insightful thoughts on Zhang et al's paper. I am also interested in your comments related to the airline companies research.
Mei Po-Kwan's writing on the Uncertain Geographic Context Problem stuck out to me in highlighting the limitations of researchers and other data consumers in understanding their data and its dynamics to the fullest. Though the author lent a significant amount of their attention to the problem of representation and explanation (Mei-Po Kwan 2012, 962), I felt that their critiques were broadly applicable. Not only does the UGCoP present itself in the interpretation of collected and synthesized information – which can often be misleading in visual representation – but it is present at the collection and design stage as well. This observation is constructive in promoting multiple-method analysis and undermining our faith in the objective, which drives us to more research and deepens our understanding of subject matter.
As is suggested by my previous statement, post-truth and post-objective approaches are extremely exciting to me; as they intertwine the typological categories discussed by MacEacren et. al. with constructivist approaches the possibilities for exploration seem truly endless. Zhang et. al. demonstrate this in their exploration of the subversive use of geotagging and social media communications, as given enough requisite data I can envision many studies of this phenomenon centering different actors. For each interest group new actions and reactions could be more completely outlined and arranged in a holistic composite.
In my own research for this course I have tried to balance in depth explorations of niche details with pattern recognition and explanation. I have also had to recognize the fact that even as I explore the uncertainty of others and attempt to explain their interactions in light of it, I bear my own lack of contextual understanding about each of them – a feature which can stimulate further exploration of my subject matter and data in the future.
I have to wonder, however, what this means for where research begins and ends; how much can one truly contribute to our collective understanding of human relations when research subjects are engaged in the construction of their own dynamic truths in a flexible social environment? It would seem to me that this is something that must be balanced out by literature reviews and case-study informed theories of human behavior at a general level. But is this not what researchers already engage in as a collective? I have to wonder whether I am unable to envision alternative constructions or whether a post-objective approach is intuitive but cognitively distant. I find the second possibility to be plausible based on the fact that our socialization in conditions of uncertainty teaches a scrutiny which contends with our more formal education in technical methodologies.
