Category Archives: cartography

That time of the semester again

This week we’ll broach the topic of datums, coordinate systems, and map projections in the GIS class that I teach at Cornell. It’s week 5+ of the semester, just enough into this stuff so that there’s some sustained knowledge growing and they now have enough of a framework onto which to hang the obvious-but-abstract-and-necessary-but-confusing-and-powerful topic.  I used to be more GIS-traditional about this stuff and dive in during weeks 2 or 3. Not any more. Much more and deeper learning taking place now that students are more confident and competent at managing and manipulating spatial data. T

Just in time, XKCD has come up with another inspired projections example to share with the class.

Citrus inspiration for cartographers

New challenges for the projection-minded. What new types of distortion can be created with these?

The Art of Ornamental Orange Peeling, circa 1905

braids and twists

braids and twists

Maps as Organizational Templates

There’s a hip trend going around, making simple maps with labeled spaces. At least one or two a week have been crossing my computer screen lately. I’ve always referred to this approach as using maps as organizational templates. In most cases the map-makers don’t go into telling a story about why the data are where they are. They’re just labeling a place with its information, and leaving the rest up to us. The map is serving as a way to represent data by virtue of its geographical location. That is, we start with some data, and that data happens to have a 1-to-1 relationship with some location, like a state in the US, or a country in the world.

We could use a spreadsheet as an organizational template instead. In fact, many of these maps started that way. Start with a spreadsheet with an alphabetical list of all 50 states (plus D.C., which often gets overlooked), and then another column in the spreadsheet has some information about each state (let’s call it an “attribute” of that State).  And maybe we know different attributes for different years.

Problem: looking at an Excel spreadsheet is boring. And it’s virtually impossible for us to envision a “pattern” from a spreadsheet. States or countries arranged alphabetically tells us nothing about the geospatial relationships among those places. Did I already mention it’s boring. Our eyes glaze over. Who wants to have glazed eyes?

Instead, by labeling each state – or country – or region – with the attribute, we can appreciate the geographic pattern of said information.  When the data are categorical or nominal, you might get a map like what the most popular boy’s name has been in each state over the last 60 years, or the girls’ names, or surnames in Europe, how the Russian language engenders the names of world countries, or what each world country is “best at” (which is a wonderfully subjective way to begin a discussion), with the label being a word or a phrase.

Such data can often be represented pictorially or through icons, like the “most famous book” in each State (again, who gets to decide that?!), or the Food of the States. At least they remembered D.C.!

I don’t know, maybe it’s just me. But I’m seeing these maps all over the place these days.

visualize El Capitan’s geology through swiping

I love it when somebody manages to collect original data of something that we’ve all seen before, with new details and insights.  This swipe-enabled image of the geology of El Capitan in Yosemite National Park is terrific.

h/t Neotorama

an idea worth supporting: an innovative, well-designed, and community-sourced Food Atlas

New to me: the current project being undertaken by “guerrilla cartographers” to create a food atlas. I love the premise, I love the process, and I know I’ll like the product. Go mappers!

Connections between hazard mapping and outcomes

Time Magazine reports on a study (pdf) that considers the connections between hazard maps and recent natural disasters.  Are “bad maps” to blame for greater-than-expected damage, death, and destruction?  Among other issues, the study authors suggest that mapmakers may lack adequate “humility and caution.”  Of course that may be true, but it’s so much more complicated than that.  Map makers rely on the data they have, not the data they want. They are required to generate maps that rank risk based on models that necessarily have fragmented, incomplete, sampled, and uncertain data.  The cartographic symbology necessary to communicate this uncertainty is often lacking.

Which for me gets to the interesting set of questions. Understanding the connections between a map maker, the representation itself, and the decision makers on the other end.  Few map makers set out to create a map that leads to poor decisions. What happens along the way?  How can we do better to reduce confusion and re-align intentions? How can we improve feedback mechanisms so that the next generation of maps is “better”?  Really, no map is inherently “bad” – so we need a better set of terms, and expectations, and practices, so that effective maps help support the best decisions.

new examples of geographically-informed art

Some lovely and gentle hand-drawn maps from Emily Robertson, followed by some woodcut ones via Bob Vila.

Maps of the United States shapes comprised of their bird feathers, by kelzuki on esty, and maps shaping other forms and objects, by Matt Cusick.

Finally, some philosophical thoughts linked to inspirational images, by Maptia.

new discovery: Jerry’s map, all 50 years and 2500 plates

I’m sorry I can’t find the first link that directed me to this, but today I took a look at a site I’d tagged to revisit, Jerry’s Map.  I’m loving it. Only map-making, geodesign-inspired, cartographically-motivated, color-exhilarated people might watch all 10+ minutes of his story, on video. Me?  I’m going to watch it again.

I love how he lets the cards direct his movements, and how he’s realized how the combinations from non-adjacent tiles are just as beautiful as art, and how he manages to find his balance between following spatially-autocorrelated rules and taking artistic liberties.

Cultivating Graphicacy While Teaching GIS?

There’s a side to working with maps and data that’s easy to overlook when we design our courses, and it falls under the heading of graphicacy.  My own quick definition of graphicacy is making, interpreting and critiquing of information in non-text form, including graphs, tables, figures, charts, AND MAPS.   I can’t remember first learning the word, but this Aldrich and Shepphard article (pdf) was one of the first that explained the concept.

I’ve come to strongly believe that graphicacy is a necessary and essential component of education.  One of the obvious reasons is how much we’re confronted with information in graphical form, such as this world population map from The Economist or the New York Times recent map of tornadoes and other natural disasters.  A good sense of graphicacy means that you are critical and creative with data, know when to question a representation, can envision alternative representations, can interpret the information and articulate its message.

The use of GIS presents numerous opportunities to develop strong graphicacy skills, but it’s definitely not an automatic outcome.  It includes the classification of data, and the cartographic design of layouts, but it goes well beyond that. It’s fundamental to how we expect to communicate with the rest of the world about what a GIS analyses means.  It’s not something that’s a separate topic to be added to a GIS course. It’s an understanding that needs to be cultivated throughout, in every lab and exercise that a student completes, and in every mapped representation that they create and encounter.  It’s the understanding of how maps complement and support learning on many levels.

GIS & the Humanities at UCSB, Day 2

The two-day mini-conference on GIS and the Humanities was sponsored by the Interdisciplinary Humanities Center at UCSB.  Day 1 included talks by archaeologists, historians, and language specialists, among others, who all have their own reasons for exploring maps and mapping.  Greater insights, new insights, tools of an evolving trade, being known for doing something different.

Early on Day 1, someone in the audience asked what the forecast and progression for these major projects was.  In traditional humanities scholarship, you do a lot of reading and research, then you write it up in a book, then it’s published and it’s the end of a particular scholarly sequence, and there is public recognition of closure.  With some very large academic GIS projects, the “end” looks different, if there is one.  Maybe it’s the end of a data set, or the end of a particular set of questions. Or maybe it’s just the end of the funding source, and then the programmers are let go or move on to something else. Or when the lead researchers move to another institution.  Or the GIS software, or the optimized browser, or the computer’s operating system, changes, and there are no funds to pay for updating.  Those are ends too.

Can the massive projects be better “chunked” – so that you don’t spend multiple years on something and still have someone underwhelmed with the results?   Will you have plans for project sustainability be part of the original proposal?  Isn’t this part of why NSF now requires a data management plan?  Also reminds me of the resistance of some departments at some universities to admitting someone directly into a PhD program, rather than a MS/MA as a stepping stone.  If something happens (and something often does), you at least have some degree in hand, instead of being 7+ yrs into something with nothing to show.

Day 2 of the conference began with my own talk, about maps functioning as both metaphors and analogies, and the complexities of supporting an argument with maps.  I discussed Reg Golledge’s representation of increasingly clearer cognitive maps as a metaphor for knowledge in general, and a relevant one for experts using new tools to communicate about their subject area to novices.

Ian Gregory’s Mapping the Lakes District project was an example I highlighted for several reasons, mostly because I like the work.  It also illustrates a point of contention: that generating surfaces (such as making a kernel density surface from point data) requires careful attention to kernel sizes, and attention to the notion of “non-uniform distribution of space.”  GIS software typically assumes that an “event” – such as a point dot indicating a location mentioned by a character in an 19th century book – “could” happen anywhere (or at least to the edges of the map extent of that project, almost always a rectangular area defined by the data set with the largest geographic extent), and proceeds to create a surface throughout, showing relative density of where the data set indicates events occurred.  But in reality most events can’t and don’t happen everywhere across a wide region.  There are reasons why Coleridge and Gray went where they did, and where they didn’t, and people don’t move around in the ways that some surficial representations suggest.

For example, if you create a kernel density surface with known zebra mussel locations (mapped within a stream channel, for example), the software will return a rectangular area extending completely beyond the stream’s reach, miles away from the water where a zebra mussel lives.  With some GIS software it is possible to use “barriers” or “masks” to limit the analysis to occur only within one area or another, but this requires steps beyond the defaults (such as setting Processing Extent parameters as an Environmental Setting in ArcGIS10).  Guess I just like points and lines better than areas or surfaces to indicate where events and movement happen. When we don’t know exactly where an event took place, and a single point in an exact spot is misleading, then a generalized surface is possible but buffers and alternative cartographic representations are also solutions.

I showed examples of how the Google Maps API is becoming a standard platform for *any* kind of project requiring simple navigation, such as Google’s fractals program, and the Google Art project.  Of course no “north” arrow needed on the fractals…  I also talked about a few relevant humanities projects at Redlands, some of which are described briefly here, plus some efforts at representing mapped uncertainty and new approaches to documenting humanities-focused metadata. I think it went pretty well. Always hard to tell, and I read as much into what people don’t say as what they do.

Fellow geographer John Agnew (UCLA) was the conference’s second keynote speaker.  John is a senior scholar who has written prolifically about political geography and notions of space and place.  He focused on regional patterns of Italian politics of the last 20 yrs, some shifting and some enduring.  He ended with some examples of how geographically weighted regression (GWR) was being used at the municipal level to tease out local variability (of Italian electoral patterns).  For me the most interesting was the final discussion about space and place.  For John, it was clear that the GWR results were getting at place – that Town A is clearly a different place from Town B, a difference that had gone unobserved when both Town A and B had been lumped together within Province Y.  Because we had changed the scale of the aggregated data (going from province down to municipality), we did have finer resolution.  But to many in the audience, the choropleth maps did little to evoke that elusive sense of place that they expect to be able to find, somehow, in a map.  They were still seeing shapes colored blue or red, albeit somewhat smaller shapes than in the previous maps.  To an expert like John, the small shapes do embody the inherent “platial” differences he knows to exist, but the nuances were largely lost to the novice audience.  I think they were looking for something that more readily evoked the experience of place.

Ruth Mostern from UC Merced was up next.  She shared details of the Digital Gazetteer of the Song Dynasty that she and grad student Elijah Meeks have produced, and talked about some of the recent influences on her thinking. These included Michael Curry’s 2005 article, Toward a Geography of a World Without Maps, plus writings in the landscape as narrative genre, including pieces by Doreen Massey and Tim Ingold.  These and other writings have clarified her thinking about how best to model history within a (geo)database. Not through administration, but through travel.  Not as structures, but as events and processes.

Highlights from the afternoon sessions included Ben Adams (grad student in computer science at UCSB) and his extractions of text from travel blogs with which he generated stylized senses of place (I really liked them; wish I could find something about it on the web to show).  I also liked the work by Marta Jankowska on slum mapping.

Mike Goodchild gave the concluding presentation, with a few remarks on the conference itself (perhaps too much worry and emphasis about the map; a possible shift to the nomothetic over the idiographic [yes, I had to remind myself what those words meant too], but a shift that is problematic with data that are increasingly resistant to generalization; perhaps this is a call to focus more on synthesis over analysis).  He talked about the realm of alternative spaces that we now study: cyberscapes framed by usage of Twitter, Facebook, and other digital social media, profiling the work of David Crandall and Matt Zook. He also reminded the UCSB community about their new academic minor in spatial studies.

Final thoughts: an interesting gathering, very worthwhile for the knowledge gained of new and innovative projects.  Some particularly relevant in my planning of our 2011 LENS Institute on Mapping Migrations.  Thanks, Ann and others, for the invitation.