Yose Widjaja, one of my students, has written a cool graph visualization system. The easiest way to get a feel for the way it works is to look at a video:
We're looking for comments, new graphs to visualize and so on. He has a blog for the project here, where you can download a demo and learn more about it.
More like this
In her recent TED talk, JoAnne Kuchera-Morin described UCSB's AlloSphere, a new project that enables scientists to literally stand inside a three-story projection of their data:
The AlloSphere space consists of a 3-story cube that is treated with extensive sound absorption material making it one…
In the same basic vein as last week's How to Read a Scientific Paper, here's a kind of online draft of the class I'm going to give Friday on the appropriate ways to present scientific data. "Present" here meaning the more general "display in some form, be it a talk, a poster, a paper, or just a…
The Free-Ride offspring are pretty sure what I do for a living is grade papers. But seeing as how they're both students, I thought I'd ask what the view of things is like on the other side of the red pen.
Dr. Free-Ride: When you come in and find me working on the weekend, what am I usually working…
I've been writing quite a bit this week about my search for a cross platform spread sheet program that would support pivot tables and make pie graphs correctly.
This all started because of a bug that my students encountered in Microsoft Excel, on Windows. I'm not personally motivated to look for…
It's all a socialist UN one-world takeover plot to get kids to "play" Data-Fighter II⢠and actually get work done, instead of wasting their lives on first-person-shooters like God intended.
It seems to look amazing but the crappy Flash video with h.263 codec used by Youtube struggles with such complex motion video. h.264 would do it better.
You may want to travel over to Resilience Alliance blog and contact them, as their work often involves visualizing connectedness of multiple systems.
I'd be interested in seeing higher res output for graphs of natural systems.
Best,
D
Tim you don't want to reinvent the wheel : in other words give it an interface to R ( r-project.org).
Very cool.
Can you visualize PubNet Data (http://pubnet.gersteinlab.org/)?
Could you use that to do a visualisation of my hard drive's directory structure, so i can easily see which folders hold all the biggest files and where i should look to go and delete stuff to free up space?
Looks intersting, tim. But what could you use it for. There already is a form of 3D perspective on car nav. that has that kind of plane approaching approaching a runway sort of visual.
Very nice! I've been interested in visualisations of large datasets like this for a long time. Can it be driven from a standard SQL platform? I have a database I'd love to navigate like this.
You are aware of this of course?
http://www.thebrain.com/
It's really hard to get a feel for what's happening without better-quality video. Could we see an mpeg or something not forced through YouTube?
Offhand though, I'd say that if you're doing 3D visualization of such primitive shapes, you really need to give better attention to depth perception. It's nice to add more dimensionality, but computers are terrible at representing 3D spaces, and we're terrible at getting our visual cues from them. So if you're relying on the 3D space to convey information (which you're doing necessarily, since it can't be "projected" back down to 2D mentally by the user), you should help us out as much as you can. For example, you use node size as a cue for some data elements, but node size scales with distance. So in order for me to accurately perceive node size, I should be able to accurately perceive distance. You've removed one of my major cues (relative size), and the graph is featureless enough that not many other cues are left. You need to add something else - distance fogging, for example.
Overall I'm unconvinced that a 3D representation is as useful as it is cool.
I'm not sure that this is useful, but it sure looks fun. How about making it into a 'cyberspace' based computer game?
Thanks for the feedback everyone! Only really had a chance to read them today.. thesis paper due soon.
> It's all a socialist UN one-world takeover plot to get kids to "play" Data-Fighter II? and actually get work done, instead of wasting their lives on first-person-shooters like God intended.
Actually, it's not out yet, but the camera system used to be a ninja.. so it *was* an FPS
> It seems to look amazing but the crappy Flash video with h.263 codec used by Youtube struggles with such complex motion video. h.264 would do it better.
There's another video, the quality is a bit better : http://www.youtube.com/watch?v=I2d312_dXEs
> Could you use that to do a visualisation of my hard drive's directory structure, so i can easily see which folders hold all the biggest files and where i should look to go and delete stuff to free up space?
> Looks intersting, tim. But what could you use it for. There already is a form of 3D perspective on car nav. that has that kind of plane approaching approaching a runway sort of visual.
Its main aim is to visually display social networks, but it actually is more powerful than just that and can definitely be used to view any graphs. DNA structure viewer maybe?
> Very nice! I've been interested in visualisations of large datasets like this for a long time. Can it be driven from a standard SQL platform? I have a database I'd love to navigate like this.
The graph display system uses a back-end that can be changed.. so this can definitely be something that's done in the future.
> You are aware of this of course? http://www.thebrain.com/
I saw that before, it would be interesting if skyrails could go in that direction..
> It's really hard to get a feel for what's happening without better-quality video. Could we see an mpeg or something not forced through YouTube?
> Offhand though, I'd say that if you're doing 3D visualization of such primitive shapes, you really need to give better attention to depth perception. It's nice to add more dimensionality, but computers are terrible at representing 3D spaces, and we're terrible at getting our visual cues from them. So if you're relying on the 3D space to convey information (which you're doing necessarily, since it can't be "projected" back down to 2D mentally by the user), you should help us out as much as you can. For example, you use node size as a cue for some data elements, but node size scales with distance. So in order for me to accurately perceive node size, I should be able to accurately perceive distance. You've removed one of my major cues (relative size), and the graph is featureless enough that not many other cues are left. You need to add something else - distance fogging, for example.
> Overall I'm unconvinced that a 3D representation is as useful as it is cool.
You have a very valid point about depth perception. Having used the system a lot (of course, I had to use it when testing it etc), you're quite right about size not conveying as much because of distance variation. However, a lot of the attribute highlighting and visual cues comes from other things, such as node color, size, edge shapes. In particular, positional bindings (e.g node Y position will represent some attributes), which would have been more constrained in 2D, is now doable in 3D.
And furthermore, 3D is a superset of 2D. Skyrails can display things both ways, and even do a "2D peek" to see the points of the graphs further away, or just use it normally like a 2D viewer. You can see this in action in the video I posted above : http://www.youtube.com/watch?v=I2d312_dXEs
> I'm not sure that this is useful, but it sure looks fun. How about making it into a 'cyberspace' based computer game?
Actually, it was one :) http://www.cse.unsw.edu.au/~wyos/cubeshatterexplo/
Thanks