Wikidata

Martin Poulter, Wikimedian in Residence at the University of Oxford, has worked on selecting and sharing datasets from Oxford's Gardens, Libraries and Museums (GLAM) as Linked Open Data in Wikidata, and creating web-based applications as proof-of-concepts to explore new ways to make GLAM's collections discoverable and visible. The short video below summarises the work Martin has undertaken and you can read more in his blog.

 

Wikidata and Oxford GLAM

https://www.youtube.com/embed/6EOlT5Y6PWw?wmode=opaque&controls=&rel=0" name="Wikidata and Oxford GLAM" frameborder="0" allowfullscreen="" id="Wikidata and Oxford GLAM

 

What Wikidata offers the GLAM Digital Strategy

The GLAM Digital Strategy has undertaken some interesting research into audience archetypes. This work examines the many different motivations and aims people can have when engaging with our GLAM institutions expressed as outcome statements, from “have fun” to “use collections in teaching”. The technology we use in GLAMs can help users in these goals, or can throw up frustrating barriers, and this strategic work explores how it could help, in particaulr to lead user-centred design.

Meanwhile, open platforms like Wikipedia continue to be the principal way in which people encounter cultural heritage. A growing “GLAM-Wiki” movement involves cultural institutions and volunteers in sharing collection data and building new tools, with some of those data sets coming from Oxford University. So what is the overlap between the user outcome statements that are shaping Oxford GLAM's aspirations and what Wikidata enables? Read more...

Expand All

 

Wikipedia and its sister-projects—collectively known as the Wikimedia projects—are among the main ways in which the public find out about art, both directly and by the influence they exert on search engine results and voice interfaces such as Siri. Often the views of a museum’s art collection on Wikipedia will greatly exceed views on the museum’s own web site. This applies to the Ashmolean, which has 51 million views per year on images in Wikipedia articles as opposed to 2 million views per year on its own site. In addition to these views, the Wikimedia projects make available data and images to third-party applications and sites.

These projects are driven by volunteer labour that is often invisible to the institution: from writing or translating Wikipedia articles, to sharing or describing images on Wikimedia Commons, the repository of freely reusable digital media. A large proportion of the effort goes into creating Linked Open Data in Wikidata, a knowledge-base that is becoming a hub for the world’s structured knowledge in the way that Wikipedia is the dominant starting-point for people seeking reference material online.

Sometimes this activity is part of a visit: having been to the Ashmolean, someone wants to share what they have learned with a wider audience. Often the motivation is broader than one institution: an enthusiast wants to document a particular artist or era of painting, and the Wikimedia platforms let them do that for works across all the relevant collections.

This article profiles some of the volunteer contributors who have built this online presence for the Ashmolean, the kind of work they do, and some of “behind the scenes” work that is public but less visible than the Wikipedia articles. It is based on information made public in user profiles and contribution records. The people mentioned are by no means an exhaustive list, but rather a set of examples.

 

Working across multiple platforms

Jane Darnell (Jane023) is an American living in the Netherlands who spends a lot of time documenting art on Wikimedia projects. Her main focus is on Dutch paintings of the 17th century; works that are strongly represented in Dutch museums like the Rijksmuseum and the Mauritshuis and also further afield, including the National Gallery, the V&A, the Fitzwilliam and the Ashmolean.

She has imported data from exhibition catalogues and from art databases, including aggregators such the Dutch database RKDImages. Listeria is a software tool that creates and automatically updates lists in a Wikipedia format. Jane uses this to create many lists related to art, (some also in Dutch) summarising the paintings in a particular collection or by a particular artist. These include paintings by Rachel Ruysch, by Adriaen Coorte, and by Clara Peeters, each of whom are represented in the Ashmolean. No less important for Jane are the lists that represent themes crossing multiple collections, such as self-portraits of women.

Among her many Wikipedia articles about paintings is “Still life with flowers on a marble slab”, which notes that one version of the painting is held at the Ashmolean. The copy is one of the paintings bequested to the Ashmolean from Daisy Linda Ward. Since the bequest overlaps with Jane’s interest in Dutch art, she has worked on documenting it, creating a category on Wikimedia Commons, uploading some public-domain images from RKDImages and also creating Wikidata representations of the different paintings.

Another Wikipedia article mostly by Jane is about Theatrum Pictorum, a 17th century art catalogue. Most of its content is a table generated with the help of Wikidata, showing one way in which structured data supports the creation and accuracy of Wikipedia.

As well as representing an artwork’s creation, materials and ownership history, Wikidata can also record what the art depicts. It already represents large numbers of artificial objects, of species, of people and of abstract concepts: a depiction is represented by a link between the painting and object. Part of Jane’s work involves adding these depiction statements. For example, after creating a Wikidata entry for Abraham van Beijeren’s “Still Life with a silver Wine-jar and a reflected Portrait of the Artist”, she added eleven “depicts” statements. As a result, this item would show up in a Wikidata query for still-life paintings that depict an oyster, or for works by Dutch artists that depict a pomegranate. Depiction statements can be very specific, so one of the depicted items Jane identified is Kraak porcelain, a type of Chinese export porcelain that was common in Dutch paintings.

Asked why she spends so much time doing this work, Jane replies:

This is an insane hobby that I love. Motivations range from “teaching by showing, not talking” to “guerilla girl edits” to “disrupt the corrupted art market” by flooding it with information.

 

The Sum of All Paintings

Maarten Dammers (Multichill) is a very active volunteer on the Wikidata Sum of All Paintings project; the ongoing effort to document all notable paintings which at the time of writing lists 365,000 works. His mother language is Dutch and he also contributes in English, German and French.

The work of Sum of All Paintings involves aggregating art data from multiple sources and presenting it in tables and image galleries. Maarten has written programs that import into Wikidata from databases such as ArtUK (formerly the Your Paintings database). He has also created Listeria lists that cover 250 museums and collections including the Ashmolean. The number of paintings listed in each collection depends on how easy it is to access and reuse that collection’s metadata, either via the institution itself or a catalogue like ArtUK. For the Metropolitan Museum of Art, the Smithsonian, and other institutions with open data policies, many, many collection items are listed.

As well as giving an overview of paintings known to Wikidata, these lists identify gaps in information: the paintings lacking an artist, location, or date. An overview table lists completeness for nine types of information: each time a painting acquires a “genre” or “depicts” property, the relevant percentage goes up. At the time of writing, the Ashmolean data set has a high level of completeness for materials, for creation date and for depiction statements, but a relatively low level of completeness of artist.

 

Documenting art in Linked Open Data

Marsupium’s mother language is German, but they also contribute to Wikipedia in French and English as well as to Open Street Map. Their main interest is representing fine art in Wikidata.

Users can propose attributes that Wikidata should represent. These are decided on in a community discussion. Marsupium has used this process to create some authority file properties for art, such as the Getty Iconography Authority ID.

Marsupium has created representations for the Ashmolean’s exhibition rooms. These exist as category tags in Wikimedia Commons and as items in Wikidata. They allow photographs to be tagged with the room where they were taken and items in the Ashmolean to be tagged with their location. A similar structure is already in place for the Louvre, making it possible to browse collections online by room or look up which room a specific artwork is in. Marsupium has also created Wikidata records for some objects that are described on collections.ashmolean.org.

Missvain’s profile just reads “All the paintings!” She has created more than eight thousand items on Wikidata, including for paintings, art collectors, and exhibitions. One of the Ashmolean paintings among these is van Dyck’s “Deposition”, for which she got details from the Ashmolean’s online catalogue, linking back to the catalogue for citations. On English Wikipedia, she writes biographies, especially of artists. She is interested in the American portrait artist Cecilia Beaux, so is using Listeria to list her works.

Edoderoo is a Dutch user who translates English labels and descriptions into Dutch. The items they work on cover all sorts of topics, including personalities, sports, and municipalities. They have used a bot to add the Dutch word “schilderij” (painting) as a description to paintings in Wikidata. Sometimes the descriptions include the artist name: for instance, “Chrysanthemums” (EA1966.161) is described in Dutch as “schilderij van Shibata Zeshin”. With Wikidata supporting around 300 languages, many Wikidata users contribute by translating names and descriptions into their own languages.

The Sum of All Paintings project lists 55 volunteers with different focuses and interests. There are more contributors who work with art data who have not signed up to that particular group.

 

Museum visitors sharing photographs

Ethan Doyle White is an amateur photographer who travels around Southern and Eastern England and elsewhere. After his visits, he uploads his photographs to Wikimedia Commons with a short description. These travels have included multiple visits to the Ashmolean, resulting in 66 photographs related to three-dimensional art and artefacts. He notes down the inventory numbers of the objects he photographs and includes those in the descriptions on Commons.

Although the descriptions of Ethan’s photographs are scant, his inclusion of the inventory numbers allows other contributors, including Marsupium (mentioned above), to look up the items in the Ashmolean online catalogue and create Wikidata entries for those objects. The photos can then connect to the Wikidata item to display information about the photographed object.

A photo by Ethan Doyle White on Wikimedia Commons, with data about the object drawn in from a Wikidata representation created by Marsupium.

John Byrne (Johnbod) has shared many hundreds of photos from cultural trips around the UK and abroad, including to the British Museum and the Ashmolean. One of his Ashmolean images illustrates the Russian article on Anglo-Saxon Art.

Like other Commons users, he tags his photographs using Wikimedia Commons’ category system. For instance, to a photo of bust of John the Baptist he has given the tags Sculptures of Saint John the Baptist and English medieval alabasters so the image appears in each list. He has created new categories, including Indian Art in the Ashmolean Museum and Medieval objects in the Ashmolean Museum, to make Commons photographs of Ashmolean objects more manageable. These tags are not (yet) structured data in the same sense as Wikidata, but they make it easier to navigate Commons’ millions of images. Categories also help to draw out themes across multiple collections. For example, Indian Art in the Ashmolean Museum is a sub-category of Collections of Art of India by Museum, which points to more than seventy different museum collections.

Andrew West (BabelStone) is another Wikimedia Commons and Wikipedia contributor, photographing museum exhibits and monuments in the UK and China and sharing the photos on Commons with descriptions and inventory numbers. His profile includes a gallery of items in the Ashmolean. Some of these photos illustrate Wikipedia articles, including the English article on the Ceryneian Hind, or the German article on the Boeotian Helmet.

Third-party uses of images and data

Wikimedia’s volunteers do not just improve the Wikimedia projects. Since those projects make their content easily available to other sites and applications, more content about an institution leads to more visibility for that institution in those applications.

Crotos is a tool for faceted browsing of artworks from many collections. It is driven by Wikidata and Wikimedia Commons: a work appears if it has both an entry in Wikidata and an image in Commons. There are presently 150,000 such works, including thousands each from the Louvre, the Smithsonian, the Rijksmuseum, and the Met. Sixty-nine items in the collection of the Ashmolean are displayed, and this will grow over time.

Back to “why?”: wiki-editing and activism

Other volunteers told me about their motivations to document art on the Wikimedia projects:

I have an insatiable curiosity about the world around me and I love to research topics from art to history. This sometimes comes with a bit of a butterfly attention span, so editing Wikipedia helps me focus on topics in greater depth and brings greater rigour to my investigations. The fact that i know that other people will benefit from my interests and my editing is an added bonus.

Over the years, I've developed strong positions rooted in feminism as well as supporting LGBT rights. I believe that actions are the best way to support those views and editing Wikipedia is a political act I can build into every day life. Adding articles about the history of women and of lgbt people is a simple yet direct way of bringing them greater visibility, which in turn can support research into more under- represented topics.

I mostly tend to focus on Scottish contemporary art, in particular female artists.

I really love arts and culture, and there is no better way to learn about it in all its diversity than to edit and research it yourself. I also do it out of a sense of activism: our world's culture really belongs to all of us, and all of it is interrelated, and I want to help to make it as accessible as possible to everyone and to put all culture in its proper (broader) context.

Wikimedia projects are so widely used and are really public platforms. Contributions there are easy to find by others, can be re-used by the broadest number of people possible, and will probably stay around for a very long time - unlike many short-lived initiatives and platforms in the cultural sector, which usually have much narrower reach, benefit only the institutions themselves, and/or often disappear after a few years.

 

A theme of these responses is of Wikimedians seeing their work as complementary to that of researchers and curators, as well as to the activities of other subject enthusiasts and museum visitors.

 

What would help this community:

  • Stable URLs. The Ashmolean’s adoption of a permalink structure based on short numerical identifiers has been helpful. There are still past links for individual artworks which no longer work. Wikimedia volunteers can help the situation by replacing these. The Ashmolean can help with this by setting redirects on its server, or by proactively updating the links.

 

  • Bulk sharing of “tombstone” data including permalinks. A dataset of the Ashmolean’s collection would help Wikimedians identify items to work on and would give proper metadata to photographs on Commons including links to the relevant online catalogue entry. For someone building a complete list of art by Rachel Ruysch or another favourite artist, it is useful to have a list of relevant artworks in the Ashmolean, with links and inventory numbers. Sharing a data set from the Jameel Centre has already raised the visibility of the Ashmolean in these platforms.

 

  • Releasing biographical details about artists. Wikidata works on the “things, not strings” principle: using unambiguous identifiers rather than by names which, like “Jan van Kessel”, might be shared by multiple artists. To help create Wikidata items for missing artists, biographical details of artists need to be publicly accessible where Wikidata contributors can access them. This might be achieved by opening up part of the catalogue, or by making a third-party biographical publication easier to access online.

 

  • Prioritising online catalogue entries for items that are already documented in Wikimedia. The Ashmolean’s online catalogue is under constant development. Some works that can be found on the Wikimedia projects lack an entry in the online catalogue. The Sum of All Paintings index for the Ashmolean is a guide to this: some paintings are visible in ArtUK or RKDImages but lack a catalogue link. If the paintings have moved to another collection, then that information needs to be shared publicly.

 

  • New interfaces: Wikimedia volunteers like to create data beyond what is given in the catalogue record, such as tagging depicted items. A catalogue record might make clear the main subject of an artwork, but a subject enthusiast can pick out other things depicted, like flora, fauna or fashion, and have the time to do it. Wikimedia Commons is acquiring a depiction-tagging function but this only applies to images that are shared on that platform.

 

Martin Poulter, April 2019

 

Delve into the Collection Explorer.

 

More about Collection Explorer

Collection Explorer is a Wikidata-driven application that enables new ways to discover objects in Oxford’s GLAM collections.

The data set

Tombstone data about items from the Jameel Centre, Digital Bodleian, the Pitt Rivers Museum, the History of Science Museum, have been shared as Linked Open Data on Wikidata. The selections have a mainly Asian focus so there is an overlap of locations and of depicted entities.

What Wikidata adds

  • Alternative forms of a name: e.g., Changshu Weng, Weng Tonghe, Songchan, and Shengyuejie are the same person
  • Names in other languages: where a collection database has one name, Wikidata may have dozens or even hundreds
  • Web address of a freely-reusable image of the person or entity
  • Further information about the person or entity, from sources including the Benezit Dictionary of Artists, Encyclopaedia Britannica, VIAF, GeoNames, Getty Union List of Artist Names, and the Oxford Research Archive
  • Connections to other entities, e.g. birth place and nationality of a person

These data appear in multiple applications that are driven by Wikidata, including:

The open data set, or different parts of it, can be visualised or downloaded from Wikidata, making possible an endless variety of different visualisations. To make the data set accessible to a general audience to support exploration and learning, we have created an application.

The application

The Collection Explorer application packages some queries in a web interface, and connects them together, so that people can navigate around a web of ideas – from depicted entities to places, to artists, to artworks – just by clicking.

It generates pages for thousands of entities: for locations, cultures/time periods (e.g. Edo period Japan, Song Dynasty China), for artists, and for things depicted in art. It deliberately doesn’t provide detailed data about individual collection items, but links through to the relevant collection record.

The application detects the language of the user’s browser or mobile device and gives entity names, descriptions and object types in that language.

The web link for each profile page is a standard stem plus the Wikidata identifier for that person. This means that other software modules can automatically links to sets of pages. This feature helps us create multiple entry points to this network. These include a map of locations, a timeline of cultures and periods, a timeline of artists and word clouds of materials and object types.

http://glam-discovery.bodleian.ox.ac.uk/

A profile page

Example of a profile page in Wikidata, here showing the historical Japanese artist, Toyokuni

Name and image

Short description and alternative names

Educational links to give context: Britannica, Benezit Dictionary, Wikipedia

Connected places or people

Art works in Oxford collections associated with this entity

These thumbnails link directly to the relevant catalogue entry for the item (e.g. Jameel Centre; Digital Bodleian). The thumbnail images in the catalogue are linked directly, not copied.

Specialist links for researchers: VIAF, ORA, EMLO, GeoNames, Getty Union List of Artist Names, and a Wiki Commons link for more related images.

Social media sharing links.


 

 

 

 

An example of a Wikidata profile page for artist Toyokuni, rendered in Japanese language

For a user with a Japanese device, some features of the interface are changed: the name and description of the entity the page is about, the related places, and the Wikipedia link.


Titles of the art works are in English by default, but their types (e.g. painting, triptych) are in the user’s language.


 

 

 

 

 

 

 

 

 

Wikidata on Wikimedia Commons

 

A screenshot of a museum artefact in Wiki Commons showing the image and associated metadata

A visitor has taken a photograph of an exhibit and uploaded it to Wiki Commons.

It was a simple change to link the photo to the object’s Wikidata record.

This lets Wiki Commons display data including the object’s date, material, and type.

The name of the museum is linked to further information, including the museum’s location and official web site.

Wiki Commons automatically links to the object’s entry in the online catalogue.

 

 

A museum in Reasonator

Finally, here you can see part of the profile of the History of Science Museum in the Wikidata-driven application Reasonator:

Part of the profile of the History of Science Museum in the Wikidata-driven application Reasonator

 

 

 

 

 

 

 

 

 

 

Martin Poulter, March 2019

GLAM Digital Showcase - Astrolabe Explorer: the power of Wikidata to explore collections

 

 

 

Delve into the Astrolabe Explorer.

 

A Global Collection of Astrolabes in Linked Open Data

An earlier blog post noted how easy it is to describe a GLAM collection item in Wikidata: it’s quicker than writing a blog post in WordPress and the resulting data are endlessly reusable. This post goes into more dtail about using Wikidata’s interface to describe items from museum collections, and announcing a new tool to browse an aggregated collection.

The Museum of the History of Science recently shared catalogue data about its outstanding collection of 165 astrolabes on Wikidata. Although Wikidata already had the power to describe astrolabes, very few had been entered, so this donation is a huge leap forward. If nothing comes to mind when you hear the word 'astrolabe', here’s an image gallery generated by a query on Wikidata. Read more...

The links below will take you to some more examples of GLAM's experiments with Wikidata:

The Sidthorp and Bauer Expedition.

Manuscript Explorer.

 

You can read a summary of Martin Poulter's work with GLAM and the tools he's created on - where else but? - Wikidata itself.

Articles and blogs are released under Creative Commons Attribution-ShareAlike (CC-BY-SA) except where stated. 

List of site pages