Inspired by this lovely visualization (which I discovered recently via the always-interesting Information Aesthetics blog) of what was the new what in 2005, I decided to bring it up to date. Here’s what I ended up with after a couple of hours of hacking. Around 160 relationships in the form x is the new y.
The layout was entirely automatic, thanks to Graphviz.
Firstly, I used Google to collect results from the query “is the new” and “* is the new *”, collecting a list like this
- Pink Is The New Blog
- Green Is The New Red
- Blood Is The New Black
- fake is the new real
- Small is the new big
- Vegetarian is the New Prius
- Quiet Is the New Loud
- …
I actually had quite a bit of manual intervention at this stage (more than I realised would be necessary), removing such results as “The Oxford English Literary History is the new century’s definitive account of …” and a few others that didn’t interest me.
Next, since the capitalization being all over the place was going to hurt the automatic graphing, I globally replaced all words with their lower-case, except where it looked more attractive not to. Then, I replaced “is the new” with ” -> “, so the file would be valid dot syntax, like this
- pink -> blog
- green -> red
- blood -> black
- fake -> real
- …
Already, I could use the `dot` and `neato` Graphviz tools, and entertained myself by checking and tidying the results still further.
But for the final render in neato I adding some extra touches to the .dot source…
digraph g {
edge [
len = 1.2
]
graph [
center = true,
start = 42
]
node [
shape = plaintext,
fontname = Tahoma
]
overlap = false
splines = true
sep = .1
pink -> blog
green -> red
blood -> black
fake -> real
small -> big
...etc...
}
And there we have it. I’d really like to make something which allows me (and you) to delve into chosen terms interactively. Seeing the interesting cluster around Facebook, even in such a small set of relationships, made me want to see what else was lurking around “* is the new Facebook”, and “Facebook is the new *”. In fact, I thought the same thing about pretty much every node in the graph. So, something else for the todo list then.
Update: I explored trails starting from Facebook (manually searching, for now) and came up with another set of data. Here are two graphs generated from that data, rendered using dot and neato respectively.
I think I prefer the (less formal looking) neato example, though there are some nice features of the vertical flow in the dot graph too.
Update: this idea later got an update for the end of 2007, and was even printed in the Boston Globe.
Wow. Very nice. We did the whole thing by handusing Illustrator of all things! In fact our compilation of terms was “manual” in the sense that we didn’t search for the term…Nice to see our project evolve through your work.
That’s really cool roo. All the network diagrams like this I’ve seen before tend to be about people’s relationships. I like how you’ve done it on something different.
I wonder if the same thing would work with years, to find styles, trends and fashions associated with each year. People often say, “that is so 1999” etc…
This is really cool. What I really find interesting are those relationships where you can’t figure out which is newer (e.g. BeBo is the new MySpace and MySpace is the new Bebo)
Nice.
this is fascinating – speaking of “x is the new y”, which is a “snow clone”, I thought you might be interested in Dave Cormier’s post he wrote for the future of education conference earlier this year – equally fascinating stuff.