James Willford. Graphing Culture. Humanities. March/April 22, number 2.
:
Thursday, March 10, 2011
article about cultural analytics research in NEH Humanities magazine
Keywords: visualization, cultural analytics, softwarestudies.com, softeware studies, digital humanities, Calit2, UCSD, NEH, NERSC, HIPerSpace, Cultural Analytics, Lev Manovich, Jeremy Douglass
James Willford. Graphing Culture. Humanities. March/April 22, number 2.
James Willford. Graphing Culture. Humanities. March/April 22, number 2.
Wednesday, March 9, 2011
New article on computational analysis of one million manga pages
Keywords: manga, comics, Naruto, visualization, cultural analytics, softwarestudies.com, softeware studies, digital humanities, Calit2, UCSD, NEH, NERSC, HIPerSpace, Cultural Analytics, Lev Manovich, Jeremy Douglass, William Huber
Article: Jeremy Douglass, William Huber, Lev Manovich. Understanding scanlation: how to read one million fan-translated manga pages. Forthcoming in Image and Narrative, Spring 2011.
Article: Jeremy Douglass, William Huber, Lev Manovich. Understanding scanlation: how to read one million fan-translated manga pages. Forthcoming in Image and Narrative, Spring 2011.
Thursday, February 17, 2011
Research on Remix and Cultural Analytics, by Eduardo Navas
Keywords: calit2, Charleston, CRCA, Cultural Analytics, Eduardo Navas, Mashups, Memes, Remix, Selective Remix, Software Studies Lab, University of Bergen, UIB, University of California San Diego, UCSD, Video, YouTube
Image: Composite image of four YouTube video remixes. From top left to bottom right appear thumbnail montages of Charleston — Original Al & Leon Style!!, Charles Style, Charleston Mirror, and Charleston Mix. Larger images of the montages with proper explanation are included below as part of this introduction to my initial research on viral videos.
As part of my post doctoral research for The Department of Information Science and Media Studies at the University of Bergen, Norway, I am using cultural analytics techniques to analyze YouTube video remixes. My research is done in collaboration with the Software Studies Lab at the University of California, San Diego. A big thank you to CRCA at Calit2 for providing a space for daily work during my stays in San Diego.
What follows is a brief introduction of my preliminary interest on video remixes and how I plan to use cultural analytics to evaluate their evolution on YouTube. My current research consists of various elements, some which I will not introduce at length today, but will only mention to contextualize the use of video grid montage as an analytical tool.
Read the complete entry at Remix Theory
Tuesday, December 21, 2010
presentation at National Department of Energy Research Center (NERSC)
"The National Energy Research Scientific Computing Center (NERSC) is the flagship scientific computing facility for the Office of Science in the U.S. Department of Energy. As one of the largest facilities in the world devoted to providing computational resources and expertise for basic scientific research, NERSC is a world leader in accelerating scientific discovery through computation. NERSC is located at Lawrence Berkeley National Laboratory in Berkeley, California." (source: http://www.nersc.gov/about/).
In the Fall 2009, we assembled a data set containing 1,074,790 unique manga pages. We then used our custom software system running on a supercomputer at at (NERSC) to analyze visual features of these pages (funded by Humanities High Performance Award from NEH Digital Humanities Office.) Currently we are working on a series of articles which discuss our methods and discoveries - as well as the larger theoretical questions around quantitative analysis of large sets of cultural visual data.
Last week we visited Oakland Scientific Facility (part of NERSC) to present our research and discuss strategies for further collaboration. The visit was organized by Dr. Daniela Ushizima. Daniela who is a member of Analytics/Visualization and Math groups in Computation Research at the Lauwrence Berkeley National Laboratory. She was responsible for getting our project going at NERSC which was no small task. Thank you, Daniela!
Photo: Daniela, Lev, and Jeremy in front of one of supercomputers at NERSC.
In the Fall 2009, we assembled a data set containing 1,074,790 unique manga pages. We then used our custom software system running on a supercomputer at at (NERSC) to analyze visual features of these pages (funded by Humanities High Performance Award from NEH Digital Humanities Office.) Currently we are working on a series of articles which discuss our methods and discoveries - as well as the larger theoretical questions around quantitative analysis of large sets of cultural visual data.
Last week we visited Oakland Scientific Facility (part of NERSC) to present our research and discuss strategies for further collaboration. The visit was organized by Dr. Daniela Ushizima. Daniela who is a member of Analytics/Visualization and Math groups in Computation Research at the Lauwrence Berkeley National Laboratory. She was responsible for getting our project going at NERSC which was no small task. Thank you, Daniela!
Photo: Daniela, Lev, and Jeremy in front of one of supercomputers at NERSC.
Monday, December 6, 2010
2008 U.S. Presidential Campaign Ads
Tara Zepel. 2010.
Article: Tara Zepel. "Cultural Analytics at Work: The 2008 U.S. Presidential Online Video Ads." In Video Vortex Reader II: Moving Images Beyond YouTube, edited by Geert Lovink and Somers Miles, 234-249. Amsterdam: Institute of Network Cultures, 2011. [pdf 2.5 MB].
Article: Tara Zepel. "Cultural Analytics at Work: The 2008 U.S. Presidential Online Video Ads." In Video Vortex Reader II: Moving Images Beyond YouTube, edited by Geert Lovink and Somers Miles, 234-249. Amsterdam: Institute of Network Cultures, 2011. [pdf 2.5 MB].
During the 2008 U.S. presidential campaigns, both Barack Obama and John McCain hosted their own YouTube channels and used the online video portals to showcase speeches, interviews, commercials and debates. The adoption of this developing medium as a political communication tool provides a unique and timely opportunity for exploring the ways in which the use of online video might have certain cultural implications – for the way we understand form and visual design and for the way visual rhetoric works on the socio-political scale of the Internet.
The visualizations that follow explore the potential connections between online video and political communication by analyzing a small sample set of commercial advertisements produced during the campaign. Highlighted here are the differences and patterns found in visual form between two subsets of the sample: 1) between ads originally designed for television broadcast and ads original designed for web broadcast and 2) between ads for Obama and ads for McCain.
__________________
Research Findings:
Sample Set of 2008 U.S. Presidential Campaign Ads:
Still frames and titles for each of the 12 campaign ads analyzed. 8 ads (4 Obama and 4 McCain) were aired on television and distributed on the web. 4 ads (2 Obama and 2 McCain) were aired exclusively online. All ads in the sample were officially produced and made available by the campaigns with the exception of D6, which was produced by the Democratic National Committee. All ads are available online on YouTube or at The Living Room Candidate, the Museum of the Moving Image's online archive of presidential campaign commercials 1952-2008.
__________________
Sample Set Visualizations:
Fig. 1. Comparison of Visual Change in Obama and McCain Video Ads.
‘Summary Images’: A superimposition of frames sampled at 12 fps for each video
Frame Diff. Line Graphs: x = frame number, y = pixel difference between consecutive frames
One way to look at aspects of visual change in the advertisements is to take a single ad as the unit of analysis. Fig. 1 shows a ‘summary image’ and 'frame difference line graph' for each of the 12 ads. Both visualization techniques depict parameters of visual change related to movement. In the summary images, the blurrier the summary image, the more the video content was changing over time. So, elements that remain discernible within the image are elements that remained in the same position throughout the ad. The frame difference line graphs are like a seismograph printout, a way of seeing the rhythm and magnitude of change over time. A spike marks a difference between two consecutive frames and taller the spike, the greater the magnitude of difference. This difference can correspond to a cut, a movement of the camera, the characters, animated text, graphics, or any other variety of visual change. Viewing these summary image and frame difference line graphs for each ad as a group allows us to see interesting patterns in visual change across the sample set:
We can also zoom out and look for patterns in visual change by taking all frames in the sample set as the unit of analysis. Fig. 2 and 3 are image maps that take regularly sampled frames (at 12 fps) from all of the ads and represent them together on two dimensions.
Fig. 2a. ‘Image Map’ Comparison of TV Video Ads.
x-axis = mean grayscale value for all pixels in single frame
y-axis = mean standard deviation for pixel grayscale values in single frame
Fig. 2b. ‘Image Map’ Comparison of TV Video Ads.
x-axis = mean grayscale value for all pixels in single frame
y-axis = mean standard deviation for pixel grayscale values in single frame
Fig. 3a. ‘Image Map’ Comparison of Obama TV ads.
x-axis = mean grayscale value for all pixels in single frame
y-axis = mean standard deviation for pixel grayscale values in single frame
Fig. 3b. ‘Image Map’ Comparison of McCain TV ads.
x-axis = mean grayscale value for all pixels in single frame
y-axis = mean standard deviation for pixel grayscale values in single frame
Notice how the distribution of frames drawn from ads produced for web broadcast is scattered, with no apparent normative trend line tying them together whereas the ads produced for television broadcast, a medium with an established history of visual development, have a distinct core. And how, when looking at the subset of ads produced for TV, the frames drawn from the McCain commercials are frames that reach the extremes of the distribution. This is closely related to the earlier observation of these same ads being more visually dynamic in terms of movement.
__________________
What Cultural Analytics Tells Us About the 2008 U.S. Presidential Campaign Ads:
If we compare patterns across a larger number of visual characteristics and condense our image analysis to sample-wide averages, keeping in mind the comparatively detailed explorations discussed thus far, we can observe general trends in the visual form of the campaign advertisements across multiple dimensions of measurement:
There is a difference between the campaigns in the visual characteristics and form they choose to broadcast on the traditional political advertising medium of television. And interestingly, the results of digital image analysis challenge our prior assumptions. If the Internet was indeed a revolutionary force in the 2008 U.S. presidential campaigns, then we might expect to see this change reflected in the visual design of campaign advertising and communication. As the numbers reveal, we do. However, the implications and parameters of the patterns that emerge from the data may deviate from our expectations about how developing medium of online video should look and should function. Based on surrounding political rhetoric, public opinion and party lines, we might expect that media team for Obama, the younger and popularly labeled more “dynamic” candidate would design commercial advertisements that showcase this dynamism. Yet, at least for this small sample set, visual analysis reveals otherwise. John McCain’s TV ads comparatively more visually radical, at least for the small-scale sample set of this study. Could there be a political message correlated with the design of the McCain TV ads that asserts ‘maverick’ in response to Obama’s message of ‘change?’ Maybe. Further, why is this pattern visible in ads produced for television and not those produced for the web? If we’re interested in the looking at the changing face of political communication , these questions might be worth pursuing.
__________________
Links:
more visualizations of the 2008 U.S. presidential campaign ads (Flickr)
The visualizations that follow explore the potential connections between online video and political communication by analyzing a small sample set of commercial advertisements produced during the campaign. Highlighted here are the differences and patterns found in visual form between two subsets of the sample: 1) between ads originally designed for television broadcast and ads original designed for web broadcast and 2) between ads for Obama and ads for McCain.
__________________
Research Findings:
- While all campaign ads in the sample set were posted and distributed on the web, those ads designed for television broadcast are visually different than ads designed for web broadcast across a number of visual dimensions.
- Comparing the television ads for both candidates leads to similarly counterintuitive observations: McCain’s TV ads are more visually aggressive and radical in visual language than Obama’s.
Sample Set of 2008 U.S. Presidential Campaign Ads:
Still frames and titles for each of the 12 campaign ads analyzed. 8 ads (4 Obama and 4 McCain) were aired on television and distributed on the web. 4 ads (2 Obama and 2 McCain) were aired exclusively online. All ads in the sample were officially produced and made available by the campaigns with the exception of D6, which was produced by the Democratic National Committee. All ads are available online on YouTube or at The Living Room Candidate, the Museum of the Moving Image's online archive of presidential campaign commercials 1952-2008.
__________________
Sample Set Visualizations:
Fig. 1. Comparison of Visual Change in Obama and McCain Video Ads.
‘Summary Images’: A superimposition of frames sampled at 12 fps for each video
Frame Diff. Line Graphs: x = frame number, y = pixel difference between consecutive frames
One way to look at aspects of visual change in the advertisements is to take a single ad as the unit of analysis. Fig. 1 shows a ‘summary image’ and 'frame difference line graph' for each of the 12 ads. Both visualization techniques depict parameters of visual change related to movement. In the summary images, the blurrier the summary image, the more the video content was changing over time. So, elements that remain discernible within the image are elements that remained in the same position throughout the ad. The frame difference line graphs are like a seismograph printout, a way of seeing the rhythm and magnitude of change over time. A spike marks a difference between two consecutive frames and taller the spike, the greater the magnitude of difference. This difference can correspond to a cut, a movement of the camera, the characters, animated text, graphics, or any other variety of visual change. Viewing these summary image and frame difference line graphs for each ad as a group allows us to see interesting patterns in visual change across the sample set:
- Web commercials are more static than commercials made for television. The summary images for the web commercials of both the Obama and McCain ads reveal more distinct textual, graphic and image content than the TV commercials of either candidate. Correspondingly, the line graphs for web commercials have a lower frequency and magnitude of visual change over time.
- McCain’s TV ads are more visually dynamic. The summary images for McCain TV commercials are noticeable blurrier than those produced by the Obama campaign. This is matched by the line graphs spiking more often and with greater intensity.
We can also zoom out and look for patterns in visual change by taking all frames in the sample set as the unit of analysis. Fig. 2 and 3 are image maps that take regularly sampled frames (at 12 fps) from all of the ads and represent them together on two dimensions.
Fig. 2a. ‘Image Map’ Comparison of TV Video Ads.
x-axis = mean grayscale value for all pixels in single frame
y-axis = mean standard deviation for pixel grayscale values in single frame
Fig. 2b. ‘Image Map’ Comparison of TV Video Ads.
x-axis = mean grayscale value for all pixels in single frame
y-axis = mean standard deviation for pixel grayscale values in single frame
Fig. 3a. ‘Image Map’ Comparison of Obama TV ads.
x-axis = mean grayscale value for all pixels in single frame
y-axis = mean standard deviation for pixel grayscale values in single frame
Fig. 3b. ‘Image Map’ Comparison of McCain TV ads.
x-axis = mean grayscale value for all pixels in single frame
y-axis = mean standard deviation for pixel grayscale values in single frame
__________________
What Cultural Analytics Tells Us About the 2008 U.S. Presidential Campaign Ads:
If we compare patterns across a larger number of visual characteristics and condense our image analysis to sample-wide averages, keeping in mind the comparatively detailed explorations discussed thus far, we can observe general trends in the visual form of the campaign advertisements across multiple dimensions of measurement:
There is a difference between the campaigns in the visual characteristics and form they choose to broadcast on the traditional political advertising medium of television. And interestingly, the results of digital image analysis challenge our prior assumptions. If the Internet was indeed a revolutionary force in the 2008 U.S. presidential campaigns, then we might expect to see this change reflected in the visual design of campaign advertising and communication. As the numbers reveal, we do. However, the implications and parameters of the patterns that emerge from the data may deviate from our expectations about how developing medium of online video should look and should function. Based on surrounding political rhetoric, public opinion and party lines, we might expect that media team for Obama, the younger and popularly labeled more “dynamic” candidate would design commercial advertisements that showcase this dynamism. Yet, at least for this small sample set, visual analysis reveals otherwise. John McCain’s TV ads comparatively more visually radical, at least for the small-scale sample set of this study. Could there be a political message correlated with the design of the McCain TV ads that asserts ‘maverick’ in response to Obama’s message of ‘change?’ Maybe. Further, why is this pattern visible in ads produced for television and not those produced for the web? If we’re interested in the looking at the changing face of political communication , these questions might be worth pursuing.
__________________
Links:
more visualizations of the 2008 U.S. presidential campaign ads (Flickr)
"Analyzing patterns in media content with supercomputers and supervisualization" / Calit2 Auditorium / Dec 3 / 4:10-4:25pm
As a part of the events around 10 year anniversary of Calit2, Lev Manovich will present a short talk about the work of Software Studies Initiative:
Calit2 Auditorium / Dec 3 / 4:10-4:25pm
title:
Analyzing patterns in media content with supercomputers and supervisualization
description:
Cultural Analytics is a new methodology for media research which takes
advantage of the next generation cyberinfratsrutcure tools being
developed at Calit2. We use digital image analysis and interactive
visualization to explore patterns across large sets of images and
video - including films, animations, video games, comics, print
publications, web sites, and other media content. The talk will
highlight our current work with 1,000,0000 manga images
to analyze the relations between visual style, genre, and audience segments.
More information: softwarestudies.com.
-----------------------------------------
Exploring one million manga images on 287 megapixel HIPerSpace on supervisualization system at Calit2, San Diego.
This photo: Jeremy Douglass (Post-doctoral researcher, Software Studies Initiative) and Florian Wiencek (Jacobs-University, Bremen)
Calit2 Auditorium / Dec 3 / 4:10-4:25pm
title:
Analyzing patterns in media content with supercomputers and supervisualization
description:
Cultural Analytics is a new methodology for media research which takes
advantage of the next generation cyberinfratsrutcure tools being
developed at Calit2. We use digital image analysis and interactive
visualization to explore patterns across large sets of images and
video - including films, animations, video games, comics, print
publications, web sites, and other media content. The talk will
highlight our current work with 1,000,0000 manga images
to analyze the relations between visual style, genre, and audience segments.
More information: softwarestudies.com.
-----------------------------------------
Exploring one million manga images on 287 megapixel HIPerSpace on supervisualization system at Calit2, San Diego.
This photo: Jeremy Douglass (Post-doctoral researcher, Software Studies Initiative) and Florian Wiencek (Jacobs-University, Bremen)
Subscribe to:
Posts (Atom)