article about Software Studies work with National Energy Research Scientific Computing Center

• 
• Tweet

A new article describes how we are working with National Energy Research Scientific Computing Center (NERSC) to pioneer use of supercomputers for the analysis of visual media in the humanities context:

Linda Vu
February 18, 2010

A Computational Science Approach for Analyzing Culture


Below are some of the visualizations which use the the results of processing of our data sets on NERSC:


All Time magazine covers for 1923-1989: 3480 total.


All Time covers for 1923-1989: 3480 total.
Covers showing people: 2583.
Covers showing women: 260.


All Time covers for 1923-1989: 3480 total.
Covers showing people: 2583.
Covers showing people of color: 306.

1,000,000 Manga pages visualization project is moving forward

• 
• Tweet

In the summer 2009 we started on our most ambitious Cultural Analytics project to date - analysis and visualization of 1,000,000 Manga pages.

Here are some initial visualizations of this data - as we create more, they will be added to our Flickr set. While these visualizations are static, we are also using the unique HIPerSpace 300 megapixel system developed at Calit2 with our custom software to explore these Manga pages interactively.


Manga categories and genres graphs

Manga visualizations




Visualization shows 1048576 Manga pages
Each point represents one page

grey points - all pages in the set
red points - all pages corresponding to a single title: Anatolia Story (artist: Chie Shinohara)

to create visualizations, we measured each page separately on NERSC supercomputers and plotted the results
X axis - brightness mean
Y axis - entropy




Data: 10 titles from our complete set of 912 titles.
Visualization shows all pages (approximately 50,000) corresponding to these titles.
X axis - grey scale standard deviation
Y axis - entropy

Note: if you click on the image above, you can download 10,000x10,000 jpeg version (scaled down from the original 44,000x44,000 visualization) - 10 MB file



Data: "Tetsuwan Girl" comic by Takahashi Tsutomu. The complete comic contains 1094 pages. This visualization is a zoomed view which omits outlires - so not all pages are seen.
X axis - grey scale standard deviation
Y axis - entropy

This visualization shows how cultural analytics approach allows us to map continuous style space of a cultural data set.




Data: "Tetsuwan Girl" comic by Takahashi Tsutomu. The complete comic contains 1094 pages; this zoomed in view shows only a small part. In the visualization, comic pages are organized left to right, top to bottom.