This is a long overdue blog post about a workshop I co-organised on the topic of exploring Cultural Collections with visualisation: The Search is Over! It was conceived by Marian Dörk, Mitchell Whitelaw and Stephen Drucker, and took place during DL2014, 11-12 September 2014 at the City University in London.
The line-up of speakers promised these two days to be exciting and it was matched by an engaged and enterprising group of participants. In this post I present a summary of the two keynotes of the first day given by Lizzy Jongma (Rijksmuseum) and Aaron Straup Cope (Cooper Hewitt).
Lizzy Jongma from the Rijksmuseum kickstarted the day, declaring that the search, in fact, has only just begun. She openly spoke about the fears cultural institutions had, and still have, about the internet. In Lizzy’s word: “We were hoping, the internet would just go away. ” Seeing that the internet had no intention of leaving anytime soon, the Rijksmuseum had to do something about it, especially because most of their collection was actually online already. Almost every work was retrievable through a Google Image search – someone somewhere already digitised van Gogh’s oeuvre before museums even thought of doing so.
This helped in convincing the Rijksmuseum that they actually had nothing to lose by sharing their collection. At least then the works would be online in high-quality reproductions, instead of snapshots taken by visitors. But simply exposing their library cataloguing system didn’t work. People don’t know what to look for if they don’t know what is there. “Enter your search term here” was the most popular search term on the Rijksmuseum website. There needs to be a better system and here at least the long wait before embracing the internet paid off, as there was now a successful example of an online collections management system: Pinterest was the inspiration for the image based design of the site and the ability for users to create their own collections. 165’000 custom collections have been curated so far, according to taxonomies that often are far remote from scholarly art history: Bubbles, Breast Feeding, Porn, etc.
The museum seemingly also gives up a source of income. Sharing the images for free meant that they could no longer charge for licensing the images. It turned out, however, that the museum actually saved money with this step, as the image licensing business was never profitable in the first place.
Copyright, a considerable factor that prevents museums from sharing their collections, was less of an issue for the Rijksmuseum. “We tried to not make copyright a problem”, explained Jongma and that they started with the public domain works. The fact that Vermeer and Rembrandt have been dead for quite a while helped. Yet there still was copyright to clear – if not for the work itself – for the reproduction of the work, as well for works whose creator was still alive. Their images initially were not included in the website, to the dismay of the artists who felt left out and demanded for their images to be included.
With copyright and licensing fees out of the way, people are now able to do anything they want with the Rijksmuseum’s collection. Including, of course, seeing it in real in the museum. Another widespread fear of cultural institutions, that their online presence would make themselves obsolete, has not been confirmed in the Rijksmuseum’s experience. On the contrary, being able to experience the works online and in many of the user-created apps and remixes encouraged a lot of people to drop by the museum, which has seen a considerable rise in visitor numbers since launching the new website.
Aaron Straup Cope
Aaron Straup Cope from Cooper Hewitt, who provided the second keynote of the day, began with a confession: “We know very little about our collection”. The statement was met with understanding rather than shock as this seems to be the reality for most, if not all museum collection.
Aaron pulled up an object description, consisting of only one word and a punctuation: “French?”. Frankly, not something that gives one a lot of confidence in the authority of an institution. Another factor which prevents museums from sharing their collections online. Shortcomings in cataloguing like this are by far no exception. Yet Aaron assures that this shouldn’t hold anyone back. A catalogue is not a representation of a museum, “We have the work to back it up.”
The goal is not to deliver the most perfect online collection, but to deliver it at all. According to Aaron probably no one has ever really seen the collection, including people within Cooper Hewitt. Making data accessible is more important, and also more achievable than making it perfect.
Accessibility however consists of much more than just exposing a search interface. Cooper Hewitt learned the same lesson as the Rijksmuseum: people don’t know what to look for. So they included a random button. “Turns out people love it.”
Visual features in general turned out to be more attractive to users than input forms – who would have thought – but that requires that there is actually something to show. Not all records in the collection have images, and not all the images that are there can be displayed. Instead of showing nothing Cooper Hewitt developed a visual language to show the reason for showing nothing – for example copyright issues – or to encode metadata visually.
Where an image is available but can’t be shown, it might still be possible to show something about the image, such as its colours. Extracting the dominant colours from all the images enabled Cooper Hewitt to provide another visual interface to the collection: search by colour has become the most popular way of accessing the collection. Its success is probably based on the same foundation as the breakthrough of the random button: Serendipity. Also, colour is a universal model for accessing information across most of the sighted population and requires no prior knowledge about the content of a collection.
Besides all this attention for the visual presentation of online collections and not the least the very topic of the workshop – exploring collections through visualisation – Aaron emphasised that not only the graphics need to be designed, so does the language. This is how Cooper Hewitt is able to deal with the issue raised in the beginning; the fact that data isn’t perfect and that this might shed a bad light on an institution.
The tone of voice of the collection’s interface makes no attempt of sounding authoritative. It’s a humble, yet witty tone that stands by the imperfection of the data: “We believe that…”, “Where quite sure that…” etc. It takes a lot for an institution to admit its faults, but hopefully the success of Cooper Hewitt’s online presence will encourage more museums to share their messy data.