Reading Time: 6 Minutes
I'm a total sucker for geographical data viz and lately I've found myself losing countless hours exploring Carto's growing community gallery of maps. Seriously, if you need some visual inspiration today head on over to their site and marvel at the beautiful intersection of geography, data, and creativity. Carto even has a freemium option, so if you're feeling particularly bold today try creating your own map. They even have sample data sets that you can play with.
So there were two maps in particular that really stood out to me which wanted to share. The first is Jill Hubley's map of tree species in New York City. The map "shows the distribution and biodiversity of the city's street trees based on the last tree census". It includes more than 50 species of trees, each represented by a different colour, and plots more than 600,000 trees across New York City's various boroughs. The end result is an utterly gorgeous map with so much detail and range it's a feast for the eyes.
I think the effort that went into creating the map speaks for itself, but here's an interesting snippet from Hubley's website on the methodology:
"The data for the map comes from NYC's open data portal, where it can be downloaded by borough... The CSV files from the portal don't include geodata, so I downloaded the shapefiles, opened them in QGIS, reprojected them to web mercator, added geometry columns to the attribute tables, then exported each table to CSV".
On the map itself you can zoom into different regions of New York City and filter the map by the various types of trees. I know absolutely nothing about tree species or biodiversity but I still found myself on the site for the better half of an hour playing with the visualization. Really interesting stuff!
The next map I wanted to share was created by the Delaware Valley Regional Planning Commission (DVRPC) which visualizes 8,340 cycling trips from 220 unique cyclists in Philadelphia, tracked by an app called CyclePhilly.
This map plots cyclists trip data and highlights the most popular routes by volume. The darker the colour, the more popular the route. The map is also tons of fun to play because you can filter by the type of trip, be it a work commute, errand, exercise, social visit, etc.
It's fascinating to see how the map changes depending on the purpose of the trip and I found it interesting to consider what can be learned from this. For example, the work commuter overlay shows the heaviest concentration of trips in the city center which suggests the majority of cyclists captured by the data reside in this region (which isn't all that surprising). But I was amazed at how some commuters travel such long distances for work as the routes for a small population of riders stretched from the city center far into the suburbs. On the other hand, the social visit overlay shows a very small concentration of cycle routes in the city centre, indicating that cyclists are less likely to ride long distances to meet friends. Perhaps this is because they prefer public transit over cycling for long trips or because their social circles are concentrated within the neighborhood they reside in. Whatever the reason, the map provides some great insight into the Philly cycle scene which could have lots of different applications (e.g. consumers who want to plan a cycle route, city planners researching where to build cycle lanes, etc).
A Brief History of Maps & Data
Visualizing geographical data, or thematic mapping, is nothing new. In fact, it was one of the first forms of data visualization to emerge and I wanted to share a few great example straight from the history books. I'm not going to provide a comprehensive timeline here, so if you're interested in a crash course on the history of data viz check out Dundas' A Brief History of Data Visualization.
So if you've never heard of Charles Joseph Minard, you've undoubtedly seen his work. I first came across it in elementary school when my 6th grade history teacher showed us Minard's map of Napoleon's failed Russian Campaign. It certainly never occurred to me at the time that I would still be talking about that map almost 20 years later (life can be strange).
Anyways, if you're not familiar with Charles Minard he was a french civil engineer who, in many peoples books, is one of the pioneers of what we now call data visualization. His map of Napoleon's Russian campaign, produced in 1869 and pictured below, is considered by some, namely Edward Tufte (a brilliant statistician), to be the greatest statistical graphic ever drawn.
The map plots 6 different types of data on a 2 axis grid (trust me, that's very hard to do effectively). The 6 data points visualized include the size of the army (thickness of the line), distance traveled (distance on x axis from left to right), temperature (vertical grid below the x axis divider), the latitude and longitude, direction the army was travelling, and location relative to date. What's insane about this graphic is that it was drawn entirely by hand. No fancy data visualization or statistical software, no computer, just good ol' pen and paper (and probably a ruler).
Another great example of creative data visualization is Minard's thematic representation of domestic cattle exports in France which he developed in 1858. The map, pictured below is notable as it's one of the earliest applications of the pie chart. This graphic visualizes 3 types of data; volume of beef represented by the size of the pie, type of beef represented by the colour and share/slice of the pie, and location. Although Minard was not the first to introduce the pie chart (that credit goes to William Playfair's "Statistical Breviary" in 1801), he pushed the limits of geographical data visualization by exploring how to represent multiple datasets in 2 dimensions
What I love about both of Minard's graphics above is how timeless they are. Both of these maps are just as beautiful and sophisticated as any data visualization created today, which is even more impressive when you consider that Minard had to hand draw everything.
Finding creative and innovative ways to communicate data is something every digital analyst and storyteller should try. Although there are certainly guidelines (or rules) you should follow when creating graphics, that doesn't mean you can't experiment and explore new ways to represent the data. And you certainly don't need to be a gifted statistician or civil engineer to be a great storyteller. There are tons of web based services (such as CartoDB) that make it easy to start creating beautiful and compelling data visualizations, so think outside the box and try making something beautiful with data. All you need is some inspiration and a little creativity!
... just try not to end up on WTF Viz ;-)