The Image Archive of the ETH Library is the largest Swiss GLAM image provider. Since the ETH-Bibliothek went Open Data in March 2015, the image archive has uploaded 60,000 out of more than 500'000 images to Wikimedia Commons. The tool Pattypan is currently used for uploading.

On the image database E-Pics Image Archive Online, volunteers have been able to comment on all images since January 2016, thereby improving the metadata. And they do it very diligently. More than 20,000 comments are received annually in the image archive and are incorporated into the metadata (see also our blog "Crowdsourcing").

However, the metadata on Wikimedia Commons is not updated and is therefore sometimes outdated, imprecise or even incorrect. The effort for the image archive to (manually) match the metadata would simply be too great. A tool does not yet exist.

Challenge

A general GLAM analysis tool that compares the metadata of the source system of the GLAMs (e. g. E-Pics Image Archive Online) and Wikimedia Commons and lists the differences. The analysis tool could highlight the differences (analogous to version control in Wikipedia), the user would have to manually choose for each hit whether the metadata is overwritten or not. Affected metadata fields: Title, date, description, ...

Automatic "overwriting" of metadata (update tool) is against the Wikimedia philosophy and is therefore undesirable. Furthermore, it is also possible that Wikipedians have made corrections themselves, which are not recorded by the image archive.

Data

• Bildarchiv Online, http://ba.e-pics.ethz.ch

• Wikimedia Commons: https://commons.wikimedia.org/wiki/Category:Media_contributed_by_the_ETH-Bibliothek

• Pattypan, Wikimedia Commons Upload Tool: https://commons.wikimedia.org/wiki/Commons:Pattypan

• Blog Crowdsourcing: https://blogs.ethz.ch/crowdsourcing/

Contact

Nicole Graf, Head Image Archive ETH Library

nicole.graf@library.ethz.ch