United Stats of America – Behind the Scenes

by Andreas Koller

My newest work is a visualisation of various demographic and US election-related datasets called United Stats of America, which I created in collaboration with Melissa Kim. We’ve been commissioned by KK Outlet, the London outpost of infamous Kesselskramer from Amsterdam, and they provided us with a large amount of datasets based on research carried out by Sarah Williams, Assistant Professor at MIT. The brief was to create infographic posters based on this data for their exhibition Mapping America. Let me explain what it’s about and how me made it.

KK Outlet asked my Vienna studio Strukt to do a poster for the exhibition, but as they were busy they handed over the work to London, where I jumped on the possibility and teamed up with Melissa Kim, a fellow student at the RCA Information Experience Design course. Within one week we developed a system to visualize relevant parameters and allow the viewer to explore the stories behind the numbers.

The stories

Our goal was to create a visualisation with a narrative aspect, and we tried to tell various stories on several levels by introducing a micro and macro scale to the data, so everybody could explore the data their own way. A first glance gives an overview, while getting up close makes it possible to compare, understand the details and make your own connections at your own pace.

We wanted to highlight specific hot topics discussed during the run-up to the elections, like women voters, unemployment by gender and turnout. I myself was especially interested in understanding and explaining the unique system of the Electoral College. The most interesting aspect, next to the fact that the winner takes it all, was the number of people per elector. There is a significant difference if you live in less populated states like Wyoming, North Dakota, Rhode Island or Alaska, where one elector is voted by as less as 189.433 people – whereas in California the number of people per elector rises to 678.945 – meaning that your vote counts 3.58 times more if you live in Wyoming than in California. This remarkable characteristic fascinated me, and the „Lines“ version shows it by changing the lengths of the lines accordingly. But this is just one of the stories that are waiting to be discovered.

Comparing the least populated and the most populated state in absolute numberes: Wyoming and California. The grey line going up to the top right corner indicates the number of people per elector.

Key for the Lines graphic

The process

Based on the datavis principle of small multiples described by Edward Tufte, our visualisation goes one step further as it introduces the use of different angles to express different values. Therefore it is easy to compare one specific parameter over the whole dataset, e.g. the area of a state is the green line going off to bottom right in each graphic. Alaska is huge, Rhode island is tiny, and a lot of people live in California – certain things really jump into your face, but there is much more to see. As a European, a lot of these facts were new to me and I learned a lot.

Developing the system of small multiples with different angled lines

We’ve arranged the stars exactly according to the US flag, and put the states relative to their geographic position. That way you’d see trends from east to west, from north to south.

Sketching the geographical position related to the position in the flag

I’ve used this technique of utilising small multiples with shapes indicating certian parameters at Strukt before, when I worked with Philipp Lehmann on visualisations for Terra Mater magazine – unfortunately, they decided to go with a simpler infographic style for their magazine. It is true, the kind of datavisualisation we chose challenges the viewers, puts more power to them and allows them to discover stories on their own, while infographics usually aim for taking the user step by step through the information. But I think it is far more interesting this way.

Sketch for the infographic in Terra Mater magazine designed and developed by Philipp Lehmann and myself at Strukt, dealing with the topic of ecological footprint. May 2012.

The title

United Stats of America – it’s is a hilarious title, isn’t it? Melissa came up with this pun and we went with it. We googled and found that there has been TV series with the same name earlier this year – apparently only of mediocre quality. But just now when trying to find out who blogged about our work, I found that TIME Magazine has had the same idea for their October 2006 issue.

The data

The data arrived in CSV format and XLS files, which were formatted inconsistently, so it had to be processed and cleaned for comfortable processing. I used Apple’s Numbers to aggregate all data into one spreadsheed, where in the end I had one line for each state.

Cleaning the data and putting it into one spreadsheet in Numbers

From the final spreadsheet in Numbers I exported a CSV file to work with Processing.

The final CSV data ready to be imported in Processing

The code

It’s all done in Processing, using the table class for reading CSV data by Ben Fry. For the final typographic details, resizing to A1 size and creating the printable PDF I used Illustrator. I’ve published a Google Doc containing all the underlying data – with even more datasets than we used, e.g. gun ownership, tax or income.

Early screenshot of the sketch in development. In this image, the position of the states is alphabetic and the size of the circle represents the population.

We are planning to make the code open source, see below.

The exhibition

KK Outlet opened the show Mapping America on November 6th 2012. Our three posters are shown next to each other, printed in A1.

Photo by Marla Ulrich / KK Outlet

The website

We’ve put together a website, where you can access the images in high resolution (4000 px wide) using Smooth Zoom by AlbanX and the Mousewheel Plugin by Brandon Aaron.

The future

We plan to release the code open-source on github to allow everyone to contribute their own visualisation based on this system and on the data. Therefore I have to clean up the code and add new datasets, for example the results of the US elections. Watch this blog for updates.

The team (and two famous artists)

Schreib einen Kommentar

Deine E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind markiert *