Data explosion in the museum space

The advent of museums digitizing their artworks into vast and rich data sets has opened up radically new possibilities for searching and studying collections, as well as working with them. What used to be flipping through binders, clustering reproductions or looking at originals in the museum archives, can now easily be done on a screen with all the globally accessible artworks at hand.
Machine learning and data processing

Recent developments in artificial intelligence (AI) enable us to comprehend digitized collections in previously unimaginable ways. Employing contemporary computer vision techniques, we can use machine learning to discover broader patterns within large data sets, such as similarities that transcend categories, continents, and cultures.
A new appreciation of the collection emerges.
A new kind of Spectatorship

All at Once builds on the historical development of collecting and exhibiting: From the Wunderkammer, to the associative hanging of artworks in 18th century salons, to Malraux’s Musée Imaginaire which investigated the impact of photography on the consumption of art, or groundbreaking exhibition architecture as Lina Bo Bardi’s glass easels at São Paulo’s Museum of Art. Even Jean-Luc Godard’s bande à part (maybe involuntarily) postulates a new way of experiencing the Louvre: by running through it in just 9 minutes and 43 seconds.

We believe that digital collections and data-driven approaches open up unknown ways of analyzing and displaying museum collections and ultimately a new kind of spectatorship.
All of a museum’s digitized artifacts are laid out in a large-scale printed grid that organizes them by visual similarity. The immersive arrangement offers viewers the possibility to experience the entirety of the collection at once, and, at the same time, to discover visual and structural patterns within.

Because of the high resolution of the printed reproduction, viewers can discern details within the displayed artifacts, even at the size of a thumbnail that would hardly be recognizable when viewed on a digital screen. The printed grid becomes a massive, yet engaging index to the collection.

Using augmented reality (AR), viewers become empowered to walk themselves through important aspects of the collection. They can scan any artifact of the grid and a digital layer immediately shows different connections between the artworks in a dynamic way. For example, the AR layer can show all artworks from a specific year, independently of its geographical origins. Or it can show all the works of a particular artist throughout the collection. Viewers can follow curated storylines around topics such as race, gender, provenance on virtual pathlines and walk through these stories in the physical space.

This grid becomes an immersive index of the collection itself. It can exist by itself or complement the museum by engaging visitors to take an unusual perspective and follow first leads given by the index. It is applicable to any digitized museum collection or might even exist as a collection of all global collections.
Departing from one artifact in the collection we’re able to meander sequentially through its entirety, one by one. In that way we can watch the entire collection flash in front of our eyes—say ten artifacts per second, 36,000 artifacts per hour, 864,000 in a day.

These meandering journeys follow different sorting logics: color, visual similarity, size. At the same time we can filter by geography, time, medium, artist, or create curated journeys along themes like race, gender, provenance, contemporarities.

As opposed to walks through conventional configurations of museum galleries, viewers can experience connections within the entire collection in very little time. This is what we hope will spark interest for pecularities of a collection and help people gain new perspectives on the importance of the museum itself.
All At Once seeks to bring digital museum collections back into the physical space of the museum building.

Furthermore, the All At Once framework can bring the collection to public spaces far from the museum building itself. In this way All At Once can reach people who don’t have easy physical access to a museum or who might be unaware of how it relates to themselves.
The Archduke Leopold Wilhelm in his Painting Gallery
in Brussels by David Teniers the Younger, 1651
The Louvre scene in Bande à part by Jean-Luc Godard, 1964
Prototype of All At Once - Index using more than 13000 images from the digital collection of the Williams College Museum of Art
A meandering journey through the digital collection of The Met Museum at 10 frames per second, New York.

1. Index

2. Sequence

3. Beyond the museum


All At Once proposes an algorithmic, data-driven approach to dive into millennia of material culture. It is a design concept and software framework that allows us to comprehend and display encyclopedic museum ­collections through machine learning technology in ways that were so far impossible.

With All At Once we are able to juxtapose items distant to each other in terms of space and time, or by art history conventions. By displaying these new configurations in immersive environments, viewers can understand and engage with the entirety of a collection beyond the screen of a computer or a smartphone. All At Once thus seeks to redefine the unique experience of visiting a museum building for the digital age.

All At Once is an independent research project by Studio TheGreenEyl. It received funding by the Knight Foundation through the New Museum Incubator New Inc.
The data used was provided by The Met Museum's Open Access Initiative and by Williams College Museum of Art in a collaboration with Micah Walter Studio.

Concept & Design: Frédéric Eyl & Richard The, Machine learning research: Agnes Chang
Journey through a collection:
sequence of 1.000 objects from The Met Museum's
digital collection sorted by visual similarity.
Results of shortest path relationship research among The Met dataset.
t-SNE map of 7,500 objects from The MET's digital collection
Screenshot of The MET's digital collection website

All At OnceAll At Once