By Nathan Yau of Flowing Data
It's always tough to pick my favourite data visualisation projects. I mean, it's a challenge to pick and rank your favourite anything, really – much depends on what you feel like at the time, and there's a lot of good work out there.
Nevertheless, I gave it a go, based on use of data, design and being actually useful. One unintentional theme, too: all of my picks are interactive, animated, or both, so click through to see them in all their glory…
The Refugee Project
The interactive visualisation by Hyperakt and Ekene Ijeoma shows United Nations estimates for refugees around the world from 1975 to 2012. There's a fine aesthetic and usability, but the call outs to events in history also make it one to come back to.
Researchers used birth and death places of "notable individuals" as a proxy for geographic cultural history. The premise is that if a lot of people move somewhere and live their last days in that place then that place must have it going on. So it's not quite just a map with moving bits, rather the connections represented by blue and red arcs, perhaps, mean something more.
We are now at a strange point where a significant proportion of web users know that their activity is tracked by advertisers, yet a large portion of that group still doesn't seem to care. Floodwatch, a US-focused Chrome plug-in by the Office for Creative Research and Ashkhan Soltani, tracks the ads served to American web users. The project (1) lets people look back at their banner history and (2) provides an opportunity for researchers to try to figure out the banner-serving black box. Maybe a few more people will care, then.
Mike Bostock does lots of great things with, and for, visualisation. Visualising Algorithms was one of those things this year. Bostock demonstrated processes and patterns in a wide array of algorithms for sorting, sampling and maze generation. The result was a useful learning resource and an appreciation for something that's otherwise a challenge to imagine for most people.
I forget where I was, but I was having a coffee somewhere and it started to rain outside. I saw people going for shelter, but I saw a couple of teens stop in their tracks to take a selfie. I thought that was strange, and then I felt out of touch. I just don't understand that culture, but Selfiecity by Lev Manovich, Moritz Stefaner and team does a good job dissecting it. The use of Mechanical Turk and face-recognition software brought your standard photo mosaic up a notch.
Most of us have heard the correlation-is-not-causation spiel by now. Sometimes, though, the spiel just feels like empty words that people don't really understand. Spurious Correlations by Tyler Vigen drives the point home and provides some fun in the process. About once a minute, Vigen's script looks for random things that correlate and then spits it out in a Tumblr-like fashion.
Ars Electronic Futurelab equipped quadcopter drones with LEDs and programmed them to fly in formations that could be controlled from the ground. OK, it's not data visualisation per se, but the prospect of flying pixels in the air is fairly amazing. More advanced city alerts? A Batman signal that doesn't rely on clouds for contrast? Extravagant air-writing marriage proposals? Oh, just imagine the possibilities.
This article is written by Dr. Nathan Yau, PhD, who runs the site FlowingData. It explores how statisticians, designers and data scientists use analysis and visualisation to explore and understand data. He's also written a couple of books on the topic. You can follow FlowingData via RSS, email, Twitter, and Facebook