Archive for category Visualizing Data
870 slices. 100,800 pixels per slice. Potentially 87,696,000 data points. Of course, we don’t need all of them. Some are background and others represent volumetrically hidden tissue. Since deep tissue tended to be the most red, we employed some color filtering. This was not enough. So, we added some basic edge detection. Within a particular slice, as the loader moves across a row, it checks the color value difference between one column and the next. If the difference is substantial, the pixel is added to the data set for display. Even so, the current implementation will run on my Mac Pro but not my MacBook. The loaded dataset is about 300,000 points, but all 87.6 million candidate pixels must be processed in loading. For stylization, on each iteration of the draw loop, we display about 1/6 of the total number of points, selected randomly on each iteration. This gives the model an animated quality and makes it feel less congested so that as the viewer approaches, he or she may peer into the model’s internal organs and skeletal system.
In my experience, learning to use a new tool – be it iPhone, graphics software or programming language – should not begin with a complete reading of the manual. Most technological devices offer a large number of features that are rarely ever used. Reading the manual from start to finished, the user may become overwhelmed. In practice, I think it is best to start with an idea for a project and learn to use the tool by exploring its features as the user feels his or her way through the project. This will expose you to the features most suitable to your projects and in the process of searching them out, the user will happen upon other features that may spark new ideas. Tufte did refer to some features of the iPhone of which I was previously unaware. Yet, having owned an iPod Touch for more than a year, I suspect that if I didn’t yet encounter these features on my own, they are probably not ones that will be of much help to me. Every so often, I am introduced to a useful feature of a tool I’ve been using for years and this always incites a small panic in me. Are there countless others? Maybe I could have saved thousands of hours. It’s a balance.
I disagree with Tufte’s assertion that the Stocks app falls is too much like a cartoon. An iPhone is used for quick checks, touching base. When I use the web browser, it is only out of necessity. I’d rather use a computer. The iPhone’s stock app gives you a market overview and let’s you take a quick reading of individual stocks. Deep analysis requires a larger screen. I do agree however that it would be nice if the weather app had a few more details – like hour-by-hour and weekly forecasts. But, that is what the developer API and app store are for.
Tufte argues that the best visualizations are in major scientific journals such as Science and Nature because they have excellent resources, the best minds, huge amounts of data and a lack of space, which forces them to make clear and concise visualizations. The “best” is always debatable, but I think his reasoning is sound. The lack of space argument is a good one. It is very difficult to convey something in a succinct manner, but when it is necessary to do so, it is often the case that the message is memorable.
I had previously seen Koblin‘s House of Cards project, which is an amazing visualization and a generous offering to the open-source data-visualization and “music video without video” community. The street scenes are particularly useful and the one of Thom Yorke is somewhere between haunting and eerie. The “video” ran impressively in my browser.
I was most intrigued by Ten Thousand Cents. This project reminded me of the artist Santiago Sierra, who amongst other things, hired day laborers to support a piece of Sheetrock at a 65-degree angle for an entire day. Despite being a representation of a $100 bill, Ten Thousand Cents does not have the heavy political overtones of Sierra’s work. The concept is neatly squared and the use of web labor seems entirely unproblematic, even healthy and worth encouraging. The sheep project did not capture my interest as much. Unlike the $100, the sheep visualization did not have a culmination, a big picture. Choosing one of the tiny images for enlargement was a bit arduous. The presence of an overarching image encourages closer inspection, while a scattered array feels like the end of the road. Ten Thousand Cents was a perfect illustration of how cheap, anonymous online labor can be used to build complex systems or visualizations. $100 is quite cheap for all that work. On the whole, I think these projects indicate a great potential for many wonderful uses and misuses of Mechanical Turk.
There seems to be the growing belief that Twitter will finally tell us who we really are, what we care about, etc. Of course, to gather this sort of information, we must learn how to programmatically distinguish what we care about from what we talk about. This is the hard part. On a certain night each year, American Idol would appear to be the nation’s chief concern, its central value. This is an illusion. It trends to the top, but that’s all it is – a brief conversational fad. Schmidt’s Social Collider is realistic about this, its mission is to “visualize how memes are created and how they propagate.” The meme, rather than the content of a meme, is the subject at hand. The Social Collider seems poised to show how a meme moves, but I wonder if there might be a programmatic approach to the consideration of why a meme moves, why particular ones take off and others fade. I enjoyed the aesthetic influence of the LHC – the visualization was beautiful. This reference seems to imply that there is a Higgs-Boson of the meme universe, a secret particle which brings matter into existence. Given this reference, we should also consider whether the Observer Effect and Heisenberg uncertainty principle apply to memes. Drawing upon the tools and the aesthetics of particle physics is a bold move and it seems to me that this comparison must be drawn to its logical ends, if it is to be utilized at all. The events of particle physics occur on an imperceivable temporal scale. How does this relate to meme instantiation and perpetuation? In any case, I found this to be a great visualization project and I look forward to seeing how it develops.
The works of Schmidt and Koblin are quite different from one another. In Koblin’s, the anonymous internet user is an active participant. Though he or she does not know the exact purpose of their contribution, they are aware of their making one. Schmidt’s project on the other hand, sends a drill into the ice – a probe into nature, which takes a reading and does not try to situate itself within the system. Social analysis and action each have their purpose and these two works serve as excellent examples of this difference. As a work of art, I was most taken with Koblin’s project because it drew upon the power of the internet to build something new, rather than to frame or contextualize what already exists. In college, I had a new media professor who urged that we keep our use of Google image search in check. Go outside, make new things. Visualization is always at risk of becoming a wholly spectator sport. In a sense it is that by definition, yet Koblin shows it may be something else too. The Heisenberg uncertainty principle does not only guide our understanding of how we look at nature, it is a force of nature in itself and therefore deserves to be studied for its own sake.
For the revised version of the Weather Visualization, I was interested in capturing the aesthetic of circles of confusion. Here are the results:
Sample video 1: Dodge City, KS
Sample Video 2: Amarillo, TX
Sample Video 3: Astoria, NY
This visualization draws upon realtime Yahoo Weather XML data to visualize the relationship between humidity and windspeed for a given ZIP code.
The following design by Josef Muller Brockmann was particularly interesting to me. The forms seem to imply a movement towards some “correct” order, a proper combination. In thinking about the realtime juxtaposition of two properties of the weather, I was interested in seeing what sorts of harmonies and discordances would naturally emerge. Though data visualizations all too often rely upon arbitrary relationships, we must be careful in assessing such a quality. Looking at the recorded interference patterns used to form a holographic image without knowledge of their being such, we might easily reason that the patterns are of no real significance. Encoded information requires decoding. For holograms, a properly aligned optical apparatus serves as decoder. But for other sorts of encoded information, the viewer’s mental faculties provide the only means of decoding.
Gotham is a geometric sans serif typeface designed in 2000 by Tobias Frere-Jones of the Hoefler & Frere-Jones foundry. The typeface was inspired by a style of architectural signage that was particularly popular in the fast-paced New York font world of the mid-20th Century. Gotham has been used in several high profile campaigns, most notably in Obama’s 2008 campaign signage. The Gotham font bundle is available by single-seat license for $300. This price seems reasonable for national campaigns but unreasonable for 5th grade book reports. Generally speaking, I think that artistic tools from software to fonts should be provided to students at low or no cost. I think this policy is beneficial to the creators as it readies a future generation of paying customers.
Matthew Carter is a contemporary type designer who has been instrumental to the transition to digital type. He designed Georgia, Verdana and many other noteworthy typefaces.
A two color visualization of world population densities as represented by nighttime lights.
Due to the large number of data points (205,947), the software has been optimized to do region-specific redrawing.
This data set is too large for web viewing. The source code and data are available here.
(Increase Processing memory to 1024MB)