At the moment, I mainly teach on the Research Masters at the University of Glasgow, which, of course, also involves teaching methods. I am very involved in debates around methods, both qualitative and quantitative, so I like to think that I’ve got quite a good overview. Last year, however, I was asked by a student how the content of my talk related to GIS. I hesistated for a moment – what did I actually know about GIS apart from having played around with Google Maps and read a few chapters here and there? Last week, I was given the opportunity to undertake five days of intensive GIS training with my colleagues Jane Drummond and Syed Ali Aqdus. The course included a mixture of lectures and practical exercises that, indeed, prompted a lot of interesting questions about geographical methods and the intersections of qualitative and quantitative methods. In addition, Jane gave a lot of mind-boggling examples that she had come across during her long and varied career, from cartography to programming.
Like geography as a whole, GIS has difficulties being recognised as a science of its own rather than just a weird composite. Having had a taste of this branch of geoinformatics, I can see why, as Jane mentioned, engineers and scientists tend to think that it is lacking rigour or a distinctive identity as a science. At the same time, like geography as a whole, its concern with synthesis to interact with the planet (and beyond) actually seems to be what makes it distinctive. These attempts at synthesis can take on some pretty curious forms that also include political decisions on how to map, some of which were relayed to us by Jane. For instance, latitude and longitude are not always fixed – they vary with the shape of the Earth (the Earth is not a perfect sphere). In addition, different countries use different shapes of the Earth for mapping purposes and even entire grid systems (the first thing we learned was how important it is to know what grid you are working in, since there is no universal one). A further level of distortion is added through the many different map projections, which give rise to entertaining cartoons such as this one that hangs in my office corridor:
At the beginning of the course, we were reminded of general mapping history and principles, such as sextant measurements. Geographical measurement is often associated with the military and especially imperialism. As Yves Lacoste put it in 1976: ‘La géographie, ça sert, d’abord, à faire la guerre’ (‘Geography, first of all, serves warfare’). At the same time, geographical measurement can, of course, be used for more beneficial purposes, as many projects show that make democratisation (of GIS and with GIS) a focus of their practice. Here, Jane had an interesting example of an attempt to make mapping more accessible, with the explicit aim to help people without a standard address (about 4 billion people in the world do not have a standard postal address). What3Words‘ ‘mission to address the world’ divides the world into 3×3 square metres, with each square being assigned a three random words from a dictionary. This makes about 57 trillion squares! I found this project thought-provoking, because, on the one hand, it puts in question our ideas of ‘standard’ (who is included/excluded; to what standards should we aspire?), on the other hand, its search for a more democratic standard led it to an unusual solution, compared to the ‘accepted’ solutions we have seen so far. This is also highlighted by the geographer, Robert Barr, who speaks at the end of this video:
Other examples of useful ‘civilian applications’ included the Glasgow City Council’s facilities locator and LIDAR mapping for biomass monitoring purposes. Especially the data collection examples (‘on foot’, aerial photography, satellite, thermal, laser, sound etc) reminded us of the diversity and messiness of geographical information that all somehow had to be brought into relation. Not only is the data itself problematic – but how do we can we process it all together? How can we ask and solve our questions with the kind of data that we have? And, how often do we not think, as human geographers, about the processing history of a map or of data that we are given?
With my design background, I also found much of the problems encountered in image processing (e.g. through Illustrator and Photoshop) mirrored in GIS. Much of the GIS exercises (primarily conducted in Arc GIS‘s Arc Map and ArcScene), involved data conversions, e.g. from tables to points, layers, features, or from vector to raster. Some conversion even involved surprisingly manual processing, akin to the lasso technique or filters in Photoshop. Jane showed us how badly data can be degraded through processing, but also how such limitations can be taken into account. Particularly drastic examples of error came from a lack of consideration for scaling. In order to make maps more readable in smaller format, landscape features have to be smoothed or simplified, and lines have to be thickened in order to make features legible. This is called ‘generalisation‘, and people are still looking for better methods of performing it. Here, Jane noted how mappers are subjected to increasing and unrealistic pressures over the time frame in which this considerable problem can be solved.
Other points of error-related amusement came from misinterpretation of aerial data, resampling and interpolation hazards, lack of terrain knowledge, confusion of data quality and model quality, and misrecognition of building features by various programmes or processing methods. Many such errors especially came out during the 3D exercises, but also in humorous thought experiments such as a hypothetical search for sea monsters (which initially didn’t take the probability for finding sea monsters into consideration). We were also reminded how more mundane human factors play into the quality and accessibility of geographical data, especially the cost in hiring cartographers or data collectors: how much time and labour can be invested in a mapping project?
Another interesting problem was posed through colouring and other aesthetic choices. While there are attempts at standardising colours and symbols for maps, there is still a great degree of liberty that can be taken to make maps look ‘appealing’, as Jane put it. One of the first things we learned in ArcMap, after getting a grip on the basics, was how to change colours and what to consider when changing colours (e.g. what might the colours represent?). For instance, when representing slope, many cartographers use a scale that starts with green and ends with white (green valleys, white mountain tops). The defaults often seemed to follow a completely different logic, leading to some rather displeasing or psychedelic renderings, which in turn led to many GOMA and ‘more-than-50 Shades’ jokes, as well as uncomfortable 90s flashbacks (especially in the distance map department – I wish I’d taken more screenshots now!). When preparing the map for presentation in the layout view, we noticed how much you could potentially direct decision-making not only through data processing, but also through seemingly innocent aesthetic choices. Here Jane had more examples of how she had experienced such situations in real life.
What I appreciated about this course, in relation to my teaching, was that I now feel able to better look for resources, such as examples and reading suggestions, for my students. The problems raised by GIS are not only exclusive to GIS, so it is nice to be able to cross-reference issues. If any readers have any particular suggestions from their teaching or research experience, I would be grateful if you could put them in the comments!