Category Archives: Mapping

Building age maps

This was the year the building age map broke into the mainstream. Or maybe I’m just saying that because I chose to use the challenge of making one as the Final Project in my intro GIS and mapping class. Nevertheless, there are increasingly many examples available. In my estimation, their mix of colors provoke a captivating aesthetic almost like a Chagall stained glass window.

The first one to draw my attention was the Dutch building age map. Covering the entire country, it includes an incredible 9.8 million buildings!

The latest one is also interesting, though a read of the small print reveals it is not strictly a building age map.


“Modal” building age

This is the area around Cheltenham and Bishop’s Cleeve where my mum lives. Here, blue indicates older housing and red is newer (an opposite scheme to the Dutch map, and one that doesn’t strictly conform to cartography textbooks–or Edward Tufte’s guidelines of one symbolic dimension for each data dimension). Beware this note however:

Important note: Classifications are an average across the local area, rather than for individual houses, therefore the colour coding on a building is not necessarily indicative of that building.

There’s also a lot of missing data where building age is not captured. Looks like the Brits still have some way to go to catch up with the Dutch!

In my class project, students created an overall map of Lexington, KY building ages as a first step. For the second step, they chose a block or hyper-local area and supplemented it with other data they collected or created. This way the project balanced a sense of structure with a sense of freedom (students are often stymied by too much freedom and revert to mundane projects such as mapping bike racks).

It’s tough to choose an example here as they were all very good (and we’ll be posting them all as pdfs on New Maps soon!) but here’s one that achieves a particularly strong sense of layout, in addition to the great use of color and legend design, as well as a pretty unique way of collecting data.


Final Project by Kelly Jackson and Ben Mills


These are done in ArcMap, not particularly known for its design aesthetic, and so are doubly impressive. On the left you can see Lexington by building age, with the bluer outer suburbs of more recent vintage. In the inset, the students have used a data source called iTreeCanopy, which allows you to crowdsource the presence or absence of tree cover (canopy) from a digital image of an area. Like the Amazon Mechanical Turk, you are presented with an image and decide for each click you make on it whether there is a tree there or not. Do this enough times, and you make a canopy layer of sorts. The end result is a comma separated file containing lat-long points and a tree designation. (For more on the methodology, see here.)

Why not just use a canopy layer for Lexington you ask? Because the available data are nearly two decades old (1998). It would be bad manners to complain about free open data without doing something about it, so hopefully in the long run we can establish partnerships with local data providers where we’ve improved the data layer, at least locally.

Lyzi Diamond on what to learn first about mapping

Great post by Lyzi Diamond on what to learn first about mapping. I was going to quote chunks of it, but do yourselves a favor and just read it.

Oh, OK here’s an extract:

In school, typically, we learn a basic formula: “When you’re faced with this problem, use this tool to solve it.” The real world is simply not like that. So much work goes into a) assessing the problem, b) determining which solution out of the possible range of solutions is best in this particular situation, c) deciding which tool is best to execute on that solution in that particular situation, and d) doing it over again because you fucked something up. I never learned about that reality in school, and I think that was a stunting factor.

But we’re a much more technically-literate society these days, so just learning the software isn’t enough. You have to understand both the theoretical principles around the tasks you’re executing on as well as the technology that’s underlying the software. For any tool you execute in ArcGIS, there is both a geometric/spatial problem that’s being solved (in a theoretical sense) as well as a database task that’s being executed (in a technical sense). Understanding both of those things is what will make you successful in the field.

It might be interesting to put together a sample syllabus/class along these lines, throwing in the “zombie GIS” approach I wrote about recently. Here’s what I put in my own syllabus:

Course Learning Outcomes
By the end of this course you will be both able to (1) identify candidate technologies for your problem; (2) identify and secure appropriate data; (3) successfully develop a solution using appropriate GIS and mapping technologies; and (4) critique and assess maps & GIS products, including your own. Additionally, the class will emphasize a key skill: (5) finding solutions to problems. By the end of the semester you should be able to not only comprehend GIS but solve problems of GIS applications and evaluate and recommend specific solutions to real-world problems.

A little ambitious perhaps!

Mapping all NYC’s taxi rides


Fascinating patterns are revealed by this unusual data set: all of the taxi rides captured by GPS in New York City–some 173 million trips. Mapbox has some maps and analysis.

Note also this warning: the data were “de-anonymized” fairly easily.

Mapping happiness again


Sorry to repeat this but I believe the map is even cooler now. Hover over each point to see the crowd-sourced information about happiness and greenery measures (collected by students n GEO 109, spring 2014 at the University of Kentucky). Click to see the StreetView image at that location (this is grabbed automagically from Google StreetView itself!)

Finally, if you click the link in the sub-title, it will take you to a nice Streetview portfolio of Lexington, Kentucky.

How mapping was reinvented in WWII

My colleague Susan Schulten has a great article in the New Republic on how mapping was revolutionized during World War II.

Drawing primarily from the classic work of Richard Edes Harrison, whose globe-spanning maps were published in Fortune magazine, she tells the story of how Harrison came to work for Fortune and produce his legacy.

As she says, his “not quite maps” were highly striking and innovative:

The most powerful of these images anticipated the perspective of Google Earth. Here Harrison reintroduced a spherical dimension to the map, focusing on the theaters of war in a way thatfor instancerendered the central place of the Mediterranean and the topographical obstacles facing any invasion of southern Europe. 

In fact, Harrison was more deeply involved in the war effort than is generally known. During the war, Arthur Robinson was head of the Map Division of the Office of Strategic Services (OSS, the fore-runner of the CIA). Due to a lack of cartographically trained personnel, at one point he had the idea of sending a small team to New York City to pick up techniques on airbrush shading from Harrison, who was living on West 48th Street. The OSS team also visited Robert M. Chapin, the Chief Cartographer at Time magazine.

Furthermore, the Department of State contracted with the American Geographical Society (AGS) to produce a series of “hemisphere maps” in 1945, who further contracted out Harrison (at the rate of $3 an hour!). Erwin Raisz, the accomplished cartographer, was also involved in this work.**

Susan has written about Harrison previously:
Schulten, S. 1998. Richard Edes Harrison and the Challenge to American Cartography. Imago Mundi 50:174-188.

My review in Antipode of her latest book is here.

**These two paragraphs are based on my ongoing and incomplete research into the map work of the OSS.

disClosure 23 now available

Last spring (2013) I had the honor and privilege to co-teach the annual Social Theory seminar, with three colleagues, Jenny Rice, Jeff Peters, and Susan Larson. The seminar is a long-standing feature of the Committee on Social Theory, which was founded at UKY in the spring of 1989 by JP Jones, Ted Schatzki, and Wolfgang Natter (see Postmodern Contentions, 1993).

The topic we proposed, “Mapping,” attracted a superb and diverse group of graduate students. Another tradition of the CST is the production of a journal, disClosure, which is totally written and produced by graduate students in CST and the seminar. There is a nice story on this year’s editors, Rachel Hoy and Christina Williams, here.

Issue 23 of disClosure on “Mapping” (2014) is now out. It contains a great selection of content, including poetry, interviews and articles from authors near (Transylvania University, Lexington) and far (Istanbul Technical University, University of Leeds). This year the journal goes solely online.

Three of our visitors to Lexington for the seminar, Neil Brenner, Swati Chattopadhyay and Derek Gregory were interviewed by graduate students, and those interviews are now available here. I would like to acknowledge and thank the students, as well as our visitors, for these interesting interviews. They go unfortunately unnamed in the title of the piece but they are: Jessa Loomis, Lindsay Shade (Neil Brenner), Sarah Soliman and Erin Newell (Swati Chattopadhyay), and Austin Crane, Sophie Strosberg and Marita Murphy (Derek Gregory).


New #cartodb possibilities (and #fulcrumapp!)

Some new mapping possibilities have been added by the geoweb mapping company CartoDB.

Both offer exciting possibilities that I’ve not seen before. Most amazing to me is the capability now to use near real-time NASA imagery as your base map, and to choose any day in the last two years. The imagery is 5-9 hours old, and you can choose daytime or night-time views.

Because of the ability to choose any particular date, you can go back to a season (for eg observing glacier extension) or event (Hurricane Sandy, late October 2012).

The imagery comes from GIBS, or Global Imagery Browse Services at NASA, and includes a selection of different products. Here’s yesterday’s “dust scores” for example (click for interactive map):



You can combine from hundreds of other layers (eg sea ice + sea surface temp + chlorophyll). (At least at NASA you can, I haven’t checked this in CartoDB.)

Second, there was an interesting tidbit showing how to dynamically pull Google Streetview into your point data. So the idea is that if you have collected data at various points, and uploaded these as a table to CartoDB, you can create a new column in the table which will pull form the Google Streetview API and provide you with a picture of that location.

Like this (click for interactive map):


Personally I think that’s cool!

There is a slight bug for me–once I get the SV imagery, I can’t add other fields, but their blog shows this is possible. I’m missing something somewhere.

Finally, just as I was writing this, Fulcrumapp announced they’ve got sharing of their maps up and running! Not only does this make sharing (ie, feeding the data into a visualization capability such as Mapbox or CartoDB) but it’s live!



Animated map of data using cartoDB

In my last post I shared a map experiment using data collected on the smartphone with the Fulcrum app and uploaded to cartoDB.

Another option in cartoDB is animation: [click here for map since WordPress won’t allow iframe embeds].

Here I’ve set the base map background to a black outline, imported a shapefile of UKY buildings, and animated each data point by the order in which it was added.

These are relatively simple to do using the entry-level membership of cartoDB (free, but limited to five data layers or tables only). But I thought I’d share them anyway since cartoDB has some design options not found in, the usual go-to for class-based projects such as this.

CartoDB and Fulcrumapp


In my Digital Mapping class this semester I’ve assigned an exercise to collect data using smartphones. In the past I might have used epicollect, but these days Fulcrumapp is the better choice: more fully featured and better supported.

From an educational point of view the downside is that it is a professional tool with a professional price. Fulcrum’s app is designed to collect data in small teams. A team leader sets up a series of question and response possibilities using drop-down menus on a website and sends out emails to the team members to join that project (or “app”). Once you log in on your phone (using a little bit confusingly the Fulcrum App, available for both iPhone and Android), you can start collecting georeferenced data. You can also take photos at each data location. This is then all synced up to the web-based project.

Fulcrum’s pricing plan is based on how many people join your team, with prices quickly increasing beyond my reach for even five members (a 1-member single user account ifs free). Luckily Fulcrum agreed to give me a time-expired account that could be shared between all members of my class (over 200 students).

My exercise asks students to collect data across campus for several variables: how happy they feel, how much vegetation is present, what the weather is like, and how noisy it is. (We made several options for each data variable.)

Here’s the student map as of today: 155 data records (it’s due next week so not all student have collected data yet). I’ve uploaded the .CSV file into CartoDB to explore visualization options on that platform, although the students have been asked to use Esri’s for the exercise. The rectangles here show data density.

This visualization uses CartoDB’s Density tool to show rectangles. This is similar to hexbins. As you change the zoom level the data bins differently, which is what you’d want. I can’t see a way to show the attribute data in CartoDB however (ie., instead of showing just where there’s more densely collected data, to show the data variables themselves such as happiness levels).

CartoDB–am I missing something here?

CartoDB has a small selection of base maps (not as many as Esri and funnily enough not including OSM. Edit: These can be added through a simple operation, see comment below. You can add a Mapbox base map however). Another thing I need to figure out is how to rearrange the order of legend items (“somewhat happy” came out above “very happy”) when I used the Category tool.

What I’d like to know next is how Mapbox, CartoDB, Gecommons (possibly being deprecated), QGIS, and other options can “fit together” in a workflow. Are they all necessary? Which is best for students? Eg., Fulcrum is good, but their pricing plan is not compatible with academia (I got in with them via Twitter!).

Also, what am I missing by not using the Fulcrum API? I suspect live streaming of data–the next step to explore? What else can the API do? I’ve never used an API, how do I use it? Maybe I can bring live data into CartoDB? (And then…?) Much to learn.

Review of Susan Schulten’s Mapping the Nation

My review of Susan Schulten’s book Mapping the Nation is now available online at the Antipode site.