Culture is Data

Lev Manovich at The Balie by Anne Helmond

By Anne Helmond (cc) non-commercial name attribution

Paradiso was enlightened last Sunday by the presence of a true Digital Media apostle: Lev Manovich, the renowned professor of Visual Arts from the University of California, San Diego, came to give a lecture on Cultural Analytics. His lecture was part of a one day conference, Archive 2020, organized by the Dutch expertise centre for e-culture, Virtueel Platform.

Manovich used the intriguing title Activating the archive or data dandy meets data mining, in which he referenced Dutch media theorist Geert Lovink who previously described the fetish of data collection by individuals and institutions. Manovich’s talk centered on the massive digitization efforts of existing cultural assets by institutions all over the world, from ARTstor to Google Books and BBC motion gallery (and even China’s CCTV). As Manovich argues, no human being will ever be able to keep track of all this data. However increasingly measures are taken in which the digital preservation of cultural assets is turning into an obligatory act (hence the fetish reference). Moreover this institutionalized digitization is accompanied since, let’s say 2005, by the rise of huge amounts of user generated content. As Manovich mentions, the number of images uploaded every week to Flickr is likely to be larger than all objects contained in all art museums of the world. This development sees a parallel expansion of the professional cultural universe. This rapid growth of a professional universe can mainly be seen in newly globalized countries foremost due to the growth of software tools which made for the instant availability of cultural news. Everyone now has access to the same ideas, information and tools: there are no more centers and provinces. Manovich even argues that the students, cultural professionals and governments in newly globalized countries are often more ready to embrace the latest ideas than their equivalents in the “old centers” of the world.

Lev Manovich by Anne Helmond

By Anne Helmond (cc) non-commercial name attribution

All in all this has lead to an explosive growth of cultural production. This has again lead to some intriguing questions and problems: “What does it mean to be a (video) artist today and what does it mean to do cultural criticism in such a world of superabundance? Before cultural theorists and historians generated theories and concepts about relative small data sets. But how can you track “global digital cultures” with billions of cultural objects? As Manovich argues, we need some new methods to track these developments in our cultural imagination. We need a new methodology for the study of cultural processes and artifacts – including cultural production, sharing and consumption. As Manovich explains, to analyze large cultural data sets of cultural information we can apply tools already employed in the sciences to analyze big data. We can create interactive visualizations and dynamic maps of large cultural data sets to find new patterns – and to generate new theoretical questions. Traditional boundaries disappear as visualization can be seen as esthetic statements about the world, so as forms of art (see for instance Stefanie Posavec’s literary organism). Forms of cultural data mining are already starting to rise up as we are slowly shifting from a world of new media into a world of “more media”. In this respect Manovich states ‘culture has become data’. This data (including media content and people’s creative and social activity around this content, i.e. social media) can be and will be mined and visualized.lrg-literary-organism-poste

Manovich explains his new methodology of ‘cultural analytics’ as the use of data mining and interactive visualizations of large sets of cultural data in the humanities context. Manovich introduced the idea of cultural analytics first in 2005 and you can find more information about this method at

Manovich argues that if you have an interesting idea today, you can be sure someone has the same idea somewhere else. Thus it makes no more sense to experience and study these single events. We need to start studying trends and patterns in culture instead of individual projects of ideas and concepts. We need to look at these projects in a larger context of global cultural production.

But how do you put this in practice? We need to represent and work with individual cultural objects and then work to larger and larger datasets. The key differences between existing work in culture visualization and Manovich’s approach, lies in the fact that most research projects now are driven by the existing data. Manovich wants to create techniques which can be used for much larger data sets. In contrast his methodology uses the computational analysis to generate new metadata. In this way one now creates the metadata around the objects, not the patterns inside. Manovich proposes to appropriate software from hard sciences and use them to look at work of arts and cultural works.

He gives examples of a few research projects he has worked on focusing on modern art. You can for example use software to analyze large datatsets of paintings. In this way a computer can ‘see’ whether a painting is realist or modernist (by measuring grayscale, particles, forms etc). Along these lines you can analyze the development of visual culture over time. By means of image processing you can describe the paintings qualitatively in terms of numbers. We can now make new distinctions on the basis of these outcomes; a trend line. In this way one generates new questions. As Manovich states, this method is not about answering old questions. Instead it offers new visions concerning the development of modernism from realism to modernism.

Manovich did a similar project which analyzed 165 paintings by Mark Rothko using a visual super computer to extract computational data. These are all examples of easier and sometimes more productive ways to look at culture. We can get a lot of data from these methods, Manovich says. Also, born digital media is highly interactive, and it is easy to record user interaction and user statistics. We can now use these techniques to ask different questions. This could be very interesting for, for instance, reception theory; we can now analyze the actual patterns of interaction with culture.


Manovich also expanded his analysis to movies, analyzing the variance in shot lengths in movies showing a “development over time”. The average shot length of feature films between 1900-2008 gives some interesting insides into the differences in cultural history in different countries, comparing France, the Soviet Union, the US etc. (in which the Russians proved most extreme or avant-garde prone with Vertov at the one extreme and Tarkovksi at the other…).

These kind of tools would also be able to ‘go around’ the canon. Where in normal science the focus is mostly on the canon, now we can do art history about larger contexts. But unfortunately it is mostly the canon that has been digitized. We should thus expand the canon in our digitization efforts, argues Manovich. But we can not archive everything…Therefore Manovich states we should archive equal amounts of ‘important’ canonical art and ‘random art’ to balance, in his words, ‘the important stuff wit the non important stuff’.

Interface design for Cultural Analytics research environment

The critique of the audience focused mainly on the problem of how one can quantify qualitative issues? For Manovich seems to propose a shift from qualitative to quantitative analysis. As Manovich replied, quite pragmatically: it is going to happen anyway, it is what social scientists are doing. With these techniques we can do more than with a simple manually descriptive qualitative analysis. For computers can analyze things we cannot: they can find similarities and differences in similar and likely objects. And in a way the question stays ‘how do we see?’ The brain is also a kind of computer, Manovich says. Do we analyze that different from a computer?

But, on the other hand, won’t we loose a sense of meaning if we analyze culture like a thing? Manovich argues that this is of course a complementary method, we should not throw away our other ways of establishing meaning. It is a way of expanding them. And it is also an important expansion, for how is one going to ask about the meaning of large datasets? We need to combine the traditionally humanities approach of interpretation with digital techniques to find out more. And again, meaning is not the only thing to look at. It is also about creating an experience. Patterns are the new real of our society.

You can find an interview with Lev Manovich held by Virtueel Platform here and an article explaining cultural analytics here.


8 comments on “Culture is Data

  1. Anne Helmond
    May 26, 2009

    Hi, I see you used my pictures from Lev Manovich on your blog. They require a name attribution to use.

  2. jannekeadema1979
    May 26, 2009

    Hi Anne! Sorry, I only added your name with the picture title. I fixed it immediately 🙂
    Thanks for the great pictures!

  3. Pingback: Post Literacy « OPEN REFLECTIONS

  4. Pingback: Remix Theory » Archivio » Notes on Cultural Analytics Seminar, December 16-17, 2009, Calit2, San Diego, by Eduardo Navas

  5. Pingback: Culture is Data | commonsourcing

  6. Pingback: Cultural Data Mines – Culture is Data by Janneke Adema | DDRRNT

  7. Pingback: Tracking Cultural Exchange | DDRRNT

  8. Pingback: Data and Culture | paigefarrowpublishing

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Open Reflections is created by Janneke Adema



%d bloggers like this: