Amazon recommendation network
Whenever you look at an item on Amazon, the site recommends related items that you might be interested in. So in a way, these items are connected by how people buy. Artist and designer Christopher Warnow uses the metaphor to create a network of Amazon products, where each node represents an item, and connections, or edges, represent common bonds of recommendations. Simply enter an Amazon link, and Warnow's software generates a network.
For example, the image above is the network for Edward Tufte's Visual Display of Quantitative Information, although Stephen Few's Information Dashboard Design seems to have more connections for some reason. My quick guess is that book's that are less niche have more connections, because when I entered Visualize This, the network was pretty small. Although I would've thought that Tufte's book would have a larger network than Few's.
In any case, the application and Processing code is free to play with. Warnow uses Gephi for network connections and grouping. Or if you don't feel like downloading a 60mb file, you can just watch it in action in the video below.
You might also be interested in Yasiv. It's a web app with a similar idea, but not quite as slick of an implementation.
How to Make a Contour Map
Using Color Scales and Palettes in R
Build Interactive Time Series Charts with Filters
Visualize This
Hey Christopher, thank you for the link to Yasiv :). I’m the author. Why do you think it is not quite as slick of an implementation? What is missing? I would love to make it better, and it’s work in progress.
Here is the same book visualized: http://www.yasiv.com/amazon#/S.....gory=Books
Cheers :)
Andrei, i think Yasiv is very slick, but in a different way. Yours offers much more info, esp. with the sidebar. The advantage of Warnow’s design is its simplicity. I can more quickly see the groups delineated when they have unique colors, and sometimes the cover of the book isn’t really of interest/distracting. I would try having an option for a simpleview version like his on Yasiv.
Keep up your great work!
Thank you, Liam!
Good point about clusters/grouping. This is something I’m working on now. As for removing images – I’m not convinced yet. Will ask more friends what they think…
Maybe I shall add more items from Amazon to the Graph to make it more interesting for analyzing?
Truly appreciate your feedback :)!
Purely anecdotal (and not completely relevant to the post), but I started reading Tufte after picking up some of Few’s books. Few seemed a bit more practical and modern while Tufte served as the theoretical/philosophical backdrop for visualization.