Visualization of large scale datasets looks darn pretty
January 7, 2013 9:18 AM Subscribe
null_sets is a new body of artwork aimed at exploring the gap between data and information. Consisting of a set of images, this project stems from our interest in glitches, code-breaking, and translation. our custom script encodes text files as images, making it possible to visualize both the size and architecture of large-scale data sets through an aesthetic lens. So if you ever wanted to see hamlet as a jpeg and find artistic merit hiding within its code, here's your chance.
A collaborative project that examined the interplay of data, information, and knowledge won the jury prize for the Distributed Microtopias exhibition at the 15th Annual Finger Lakes Environmental Film Festival. The University of Tennessee's Remote Data Analysis and Visualization Center (RDAV) collaborated with University of Tennessee Knoxville artist Evan Meaney on the project, called Null_Sets, which is a collection of artwork that visualizes the size and structure of data. The artwork was created using an open source script developed at RDAV in which whole bodies of text, from classic literature to HTML to genomic data, can be exported as digital images. Null_Sets uses encoding to represent the changes in pixel color and intensity, and might be adapted to explore how values in a data set change. "This project makes it possible to visualize both the size and architecture of large-scale data sets through an aesthetic lens," Meaney says. RDAV's Amy Szczepanski notes "the techniques we developed in this project laid the groundwork for a larger project that will likely use the Nautilus supercomputer (managed by the National Institute for Computational Sciences) in the future."
A collaborative project that examined the interplay of data, information, and knowledge won the jury prize for the Distributed Microtopias exhibition at the 15th Annual Finger Lakes Environmental Film Festival. The University of Tennessee's Remote Data Analysis and Visualization Center (RDAV) collaborated with University of Tennessee Knoxville artist Evan Meaney on the project, called Null_Sets, which is a collection of artwork that visualizes the size and structure of data. The artwork was created using an open source script developed at RDAV in which whole bodies of text, from classic literature to HTML to genomic data, can be exported as digital images. Null_Sets uses encoding to represent the changes in pixel color and intensity, and might be adapted to explore how values in a data set change. "This project makes it possible to visualize both the size and architecture of large-scale data sets through an aesthetic lens," Meaney says. RDAV's Amy Szczepanski notes "the techniques we developed in this project laid the groundwork for a larger project that will likely use the Nautilus supercomputer (managed by the National Institute for Computational Sciences) in the future."
“In a gallery, we can analyze these data sets side by side and consider the differences between, say, Moby Dick and an X-chromosome,” Szczepanski said.
At first this project sounded like it could be interesting, but it doesn't seem as if they are actually doing any analysis, they're just programmatically translating large sets of unrelated data into large fields of color because the results are aesthetically pleasing.
posted by oulipian at 10:26 AM on January 7, 2013 [5 favorites]
At first this project sounded like it could be interesting, but it doesn't seem as if they are actually doing any analysis, they're just programmatically translating large sets of unrelated data into large fields of color because the results are aesthetically pleasing.
posted by oulipian at 10:26 AM on January 7, 2013 [5 favorites]
This kind of thing is more just random images generated from arbitrary files, which was kind of cool 15 years ago, but now is a little behind the curve.
If you want actual visualizations of large data sets, information aesthetics.
posted by signal at 10:39 AM on January 7, 2013 [7 favorites]
If you want actual visualizations of large data sets, information aesthetics.
posted by signal at 10:39 AM on January 7, 2013 [7 favorites]
At first this project sounded like it could be interesting, but it doesn't seem as if they are actually doing any analysis, they're just programmatically translating large sets of unrelated data into large fields of color because the results are aesthetically pleasing.
This is pretty much the problem with all Data Viz, as it encourages surface interaction with data (be it censuses, literature, or spreadsheets). I'm sure there is still good stuff being made in this field, but it is swamped by the amount of glossy pap that can be created by clicking buttons.
posted by The River Ivel at 10:47 AM on January 7, 2013 [2 favorites]
This is pretty much the problem with all Data Viz, as it encourages surface interaction with data (be it censuses, literature, or spreadsheets). I'm sure there is still good stuff being made in this field, but it is swamped by the amount of glossy pap that can be created by clicking buttons.
posted by The River Ivel at 10:47 AM on January 7, 2013 [2 favorites]
My data file looks like Request Entity Too Large
... the amount of data provided in the request exceeds the capacity limit.
Also surprised there isn't an option to upload a file.
posted by armacy at 10:51 AM on January 7, 2013
... the amount of data provided in the request exceeds the capacity limit.
Also surprised there isn't an option to upload a file.
posted by armacy at 10:51 AM on January 7, 2013
This is just lazy. They could at least take these images and run them through Metasynth to turn them into music. Then visualize that music with iTunes music visualizer. Then capture the visualizer output and integrate it's data with an environmentally generated third axis of data (or some other interesting data source) and then print it out on a 3D printer. Then take the 3-D object and scan it with a Kinect and then map the 3D data set onto a life size model of David Foster Wallace. But noooo, that was too much work, apparently.
posted by Phyllis Harmonic at 11:35 AM on January 7, 2013 [4 favorites]
posted by Phyllis Harmonic at 11:35 AM on January 7, 2013 [4 favorites]
If you want actual visualizations of large data sets, information aesthetics.
Interestingly, null_sets is currently the first link on that page.
posted by dubold at 12:26 PM on January 7, 2013 [2 favorites]
Interestingly, null_sets is currently the first link on that page.
posted by dubold at 12:26 PM on January 7, 2013 [2 favorites]
I'd be far more interested in stock market trade history rendered as sound. Anybody doing that?
posted by flabdablet at 5:20 PM on January 7, 2013
posted by flabdablet at 5:20 PM on January 7, 2013
« Older Houndton Tabby | Secret and Lies of the Bailout Newer »
This thread has been archived and is closed to new comments
posted by alby at 9:57 AM on January 7, 2013