Computers, Society, and Nature

I often find papers like this can be very theoretical and I can struggle to grasp the real crux of the research, but I found that Elwood presented some fairly theoretical topics clearly by providing examples. [something about the other paper]

As spatial components of new tools and technology become increasingly ubiquitous, I find that there is now an expectation for data to be presented spatially, and some disappointment when it’s not available. These papers made me think about how I expect spatial information to be easily available and consumable for me. For example, just discussing where to go skating with some classmates, we were irritated with the idea of having to use a website that just listed the rinks with their general location (http://ville.montreal.qc.ca/portal/page?_pageid=5977,94954214&_dad=portal&_schema=PORTAL) and were thrilled to find the alternative which provided the skating rinks mapped across the island (http://www.patinermontreal.ca) – to be honest, I was fully expecting a mapped version to be available and would have been shocked if there wasn’t.  Apartment-hunting without Padmapper or similar is pretty miserable, since “where” is usually the most important variable in any potential home. Clearly non-GIS users have grown to have a higher level of spatial literacy with products from Google and spatial functions in technology, but they are also coming to expect consumable spatial information to be available.

-Kathryn

This paper raises some interesting points about what interfaces are trying to do – it might seem like they are just trying to create an attractive user experience, but in an ideal world they should allow a user to learn the software through “trial and error” and ultimately come to the correct conclusions about how the software functionality works. Seems like high expectations, but clearly useful to design software for the people who will be using it!

This paper was written a fair time ago. I wonder how many GIScientists and users of GIS have continued to find the developments/improvements in GUI for GIS software fairly unsatisfactory, and turned or returned to spatial analysis through programming? I almost never touch an ESRI product anymore – I’d usually rather work through a problem in a combination of postGIS and R. Yes, partly it’s for automation, but to be honest I do find the software can be unintuitive sometimes. However, teaching postGIS to a computer programmer with no background in GIS (with any “S”) showed me that his grasp of databases and analysis were clearly strong but the lack of visualization really did hinder understanding of the analysis tools and the concept of projections – even in someone with a highly technical background. While programming is incredibly useful in GIS and spatial analysis, the visualization aspect is still crucially important too, even for learning. And (maybe because I was trained visually) I never fully trust my results until I “sanity check” them in a visual way.

Random thought – the “mapping” problem is very interesting and of course relevant to GIS, and it seems like the touchscreen has changed this a lot since it’s so intuitive. I’ve noticed several times anecdotally that very young children can work iPads so early as age 2, while they lack the ability to work a mouse on a regular computer for years. I’ve always assumed it’s an issue of dexterity but maybe it’s more to do with this “mapping.”

-Kathryn

Rinner, from the beginning, claims that user-generated information is not considered a “serious” pursuit. To test his theory, he scrapes user-generated text for geographic references, only to retroactively apply a geo-reference to the post. The researchers, who are responsible for geographically referencing the post, strip away the users involvement in the geographic context. This can introduce a large source of error into the data. Consider, on page 14, a marker’s placement corresponds to the “label in Google’s “map” view.” In the event that a user refers to a general area, choosing to place a label greatly reduces the dimension of the users perspective. Much like a home is not a coordinate, but a structure, and the surround property, land, or neighborhood. Had the researchers allowed the users to actively tag locations to their posts, they would have gained far greater insight into the users intentions.

The creation of a Master Plan cannot be accomplished using the scientific method. Studies can find correlations, in certain areas, but many fall apart when the study area is changed. This may be due to social and cultural differences. In that light, and working on the basis that people are subjective creatures, a crowd-sourced master plan can gain from further empowering the constituents. In the case of the Ryerson Master, providing the user with the ability to choose the location of their post’s geo-tag may have developed another dimension for the study.

As a side note for crowd sourced urban planning, I do not think that even high-scale maps are sufficient. The city is rarely experienced from the air. With that in mind, I am in favor of developing forum in which people can experience the city from the ground. To my knowledge, no such technology yet exists.

AMac

In their paper on the uses of Web 2.0 to support spatial decision making, Rinner et al. address one of the problems that M. C. Er identified in DSS 20 years earlier: group decision making. By using PGIS as a source of data, making decisions for a group of people is made easier. In order to test this, the authors designed Arguumaps, where users could make geographically referenced comments about campus/city life or the Ryerson identity. Though some limitations were present, the power of PGIS is clear in its application of online mapping.

Whether this data is useful in DSS is a whole other questions. The authors argue that this case study shows that it could be a useful tool by having users vote on preferred geographically reference locations. For example, instead of posting comments about their favorite restaurants, users could cast votes on restaurants over a whole array of criteria, thus helping other users pick which restaurants they would want to go to.

Presently, 5 years after this article was written, google implemenets these services on their maps. Great strides have been made in mapping as a support to spatial decision making however much ambiguity still exists over the exact definition of DSS.

Pointy McPolygon

The Rinner et al article explores the intersection of spatial decision support systems (SDSS) and volunteered geographic information (VGI) with their development and piloting of an argumentation mapping web 2.0 application intended to solicit and map spatially-grounded citizen input and discussion. 

Rinner et al single out planning processes as areas where such applications have potential.  By tying user discussions and feedback to explicit locations, the responses can serve as a qualitative gauge by decision-makers as to the relative importance of criteria within a SDSS model.  There is even the opportunity to add a quantitative element to such conversations by integrating a positive/negative rating system to the threaded messages (like a spatial Reddit!). Overall, Rinner et al’s project aims to find effective ways to crowdsource planning decisions and amplify the ideas of citizens using emerging web 2.0 technology.

Argumentation mapping and other geospatial mashups’ applicability to decision support are not without their concerns. As we learned the hard way in GEOG 407, a #neogeoweb 2.0 application is only as good as its programmer, and can only ever be as good as the underpinning API.  Four years after this paper’s publication, many of Rinner et al’s suggestions for improvements to the API are still unaddressed by Google.  Formal integration of the data collected in ArgooMap with SDSS models is still a long way off: for now, it is limited to qualitative uses.  In addition, these digital consultative avenues, even if they are improving in terms of end-user functionality, may still be exclusionary: from the digital divide, to gender dynamics on the internet, to the less affluent simply having less free time to invest in online civic discussion, there is a role for critical geographers and GIScientists to suggest ways for SDSS and its related tools to be more inclusive and thus more likely to be truly democratizing the decision-making process.

-FischbobGeo

What is the role of GIS in supporting decision-making processes? It looks like I’m not the only one left with that question after the readings…

There is the GIS tool that continuously allows to add information (the layer cake), store and organize a tremendous amount of data, analyze the data, make maps and better inform decision makers. Is it the only role? Is it enough? Does having a lot of information allows decision-makers to make better decisions? I don’t think so. As M.C Er puts it : “decisions are based on personal experience and subjective judgment.” Taking decision is then a very individual process. As sidewalkballet argue in this blog, maybe the GIS way works for a certain cognitive style of people.

There has to be something more than relying on great good more and more information to make good decisions. Is GIS playing a role in decision-making processes because it’s more cost-effective? That is really hard to answer as M.C Er points out. If we don’t know if there is some economic benefits in using GIS why is it used? Well maybe because it’s a good way to legitimize the process and legitimize the decision. Decision-makers and managers can say :” Look we are so up to date; we are using GIS!”. It can become a way for the agencies to promote ‘good practices’.

Furthermore, I think that seeing the role of GIS in a technocratic way can lead to focus more on the processes of using information (gather information, organize the information, analyze information, communicate the results…) rather than on solving problems. That could be a very comforting thing when it’s complex to identify the problem that you’re dealing with (unstructured problems). However, I don’t think that it’s the way to optimize GIS and decision-making processes.

Even new technologies and Web 2.0 can serve the purpose of simply adding information produced by citizens about local knowledge. However, as it was already mentioned in this forum, the decision makers are not necessarily willing to take that information into consideration.

What if the role was to improve collaboration between stakeholders? Rinner et al. talk more precisely about supporting deliberation in the decision-making processes rather than the decision itself. Now that is GIScience! It is about interactions between the technology, users and producers in a specific context. The outcome is a debate on how different persons view their environment. In the example of the Argumap, the thread structure reveals the spatial thought processes of participants with the relations created between arguments. I might be a little optimistic, but the debate and discussions that are created trough the interaction have a potential to reshape existing structure between agencies and citizens.

S_Ram

M. C. Er, in his paper on Decision Support Systems written in 1988, critically examines the future of a technology that – to say the least – had a lot of room to grow. While his predictions for the future may have read like a sci-fi novel at the time of its writing, they weren’t too far from the truth, especially considering the amount of technological development that has occurred over the last 25 years. While talking about considering personality and cognitive style in decision making may seem like a given to us now, the fact that he was able to recognize that need at the time is somewhat amazing.

Much of the article is still applicable today. While technologies have advanced, the exact definition of a DSS is still unclear. On the other hand, M. C. Er states that the DSS should be centered on the problem solver, while in reality this paradigm is shifting. PGIS means that the data will start coming straight from the source, enabling a more accurate use of a DSS. Facilitating the problem solver is no longer the main concern for DSS, however data quality and data sources have come into the spotlight in a world that seems exponentially more complicated than it was in the year this article was written.

Pointy McPolygon

Densham’s article reviews work on spatial decision support systems (SDSS), conceptualizing them as distinctive from the GIS of the time.  Crucially, Densham problematizes the direct application of GIS to solving spatial decision problems, proposing instead that dedicated SDSS software incorporating specialized architecture and a modular code repository or model base management system (MBMS) be developed instead. Nonetheless, the author envisions a continued role for geographers in the informed decision process, in order to avoid “the selection of variables with inappropriate levels of resolution and geographical extent… ultimately result[ing] in solutions that are deemed unsatisfactory when evaluated in terms of the quality of the decision-making process that generated them” (p. 403).

Densham argues that “[c]urrent GIS fall short of providing GIA capabilities” because “their support of analytical modeling is lacking”, their display and reporting capabilities are limited, and they “are not flexible enough to accommodate variations in the either the context or the process of spatial decision making” (p. 405). There is no date on the paper, but its most recent citation is 1991, suggesting that when this article was published, the Soviet Union was still a thing and Apple products were at their first peak of popularity: has GIS today overcome the limitations of Densham’s era?  I would argue it has. For starters, hardware and operating systems have made huge advances since the year of my birth, supporting greater data storage, graphical capabilities and computational power for analytical programs (even interpreted ones!).  Meanwhile, the advent of the internet along with new, more user-friendly scripting languages like Python, has made implementing model capabilities within GIS’ DBMS framework and the existence of code libraries—both dismissed by Densham as not feasible—possible.  These technological advances, along with improvements in graphical representation and user interfaces, have enabled GIS software to integrate decision support modules (such as the classic SDSS problem, location-allocation) directly into the software or as a plugin while retaining the potential for customization of decision models.  Today, producers of GIS software packages aggressively market their products’ SDSS functionality.

GIS and its enabling technologies have made strides, but so have competing technologies in business information/intelligence systems/software (BIS) such as the impressive Tableau software package. Despite the increasing role of spatial data visualization, analytics and decision support in BIS, development of these tools tends to be the realm of computer scientists and not geographers. Even though tech has advanced considerably across the board, Densham’s argument that geographers are not obsolete still resonates today. As geographers, we need to assert our place at the spatial decision support table, both by advancing GIScience such that GIS remains relevant to (and if possible, ahead of) contemporary decision support analytics, and by reminding software developers and decision-makers of the importance of a nuanced understanding of spatial concepts and considerations.

-FischbobGeo

I think it’s interesting that this article (written in 1988) starts the discussion by suggesting that there has been no consensus on defining what a DSS is – it reminds me of the intersection we’re at with GIS; it is a tool or a science? M.C. ER notes that some have tried to different a DSS from a non-DSS through the intention of the design, though many counter examples arise (M.C. ER’s, 1988). Similar to our discussion about GIS as tool, science, or somewhere in-between, perhaps we can use a similar methodology in how we define what a DSS is. How the DSS (or tool of a DSS) it is used, and the intention of the tool in that particular context dictates whether or not it a DSS.

M.C. ER acknowledges that “it cannot replace upper-level managers in decision making”  (M.C. ER’s, 1988), yet I would suggest that if the decision to be made involves any type of complexity, or exceptions, the group should be broadened to encompass a wider range of people. After all, it is a “support” system, implying that it is there to help with the process, rather than making decisions unless the path to a decision can always be deduced to a binary logic.  Since 1988, DSS has significantly evolved and grown in complexity and sophistication. Group DSS (I think of group DSS in terms of electronic meeting systems, such as web conferencing from around the world) have made its mark in the workplace, while I’m less sure about how AI DSSs have progressed since the 90’s. Perhaps a missing, or unsuspected trend that emerged was the integration a spatial component into a DSS, whether it be the ability to web conference from around the world, or the ability to seek new patterns that can affect decisions when factors and information can be geographically tagged.

-tranv