Wikimedian in Residence for the National Library of Scotland | 7/31/2014
files, meaning that they were uploaded as different versions of one file rather than as the separate files that they should have been. This was fixed by adjusting the metadata, and alerted us to the problem of generalised/duplicate filenames.
Procedure – GWToolset Uploads
The procedure was developed to upload digital content from the Library’s Digital Object Database (DOD) to Wikimedia Commons using the GWToolset which is still under development. Because the toolset is in Beta and the documentation surrounding registration and use of the toolset is still scattered, the procedure will be outlined in some detail here.
First, access was requested to the GW Toolset on the Wikimedia Commons Beta Cluster by writing to the Glamtools Mailing List (glamtools@lists.wikimedia.org). This allowed me to test uploads on the Beta site (where 'breaking' anything was less of an issue) before going live. Before testing, I identified metadata fields required for the various MediaWiki templates (using Template:Photograph and Template:Artwork for the first batches), and created a .json metadata mapping page to map NLS .xml metadata fields to the templates. This can also be done manually during the mapping process and then saved for future reuse. The only metadata fields created specifically for the Wikimedia Commons uploads were "Wiki_Permissions", in which the necessary license templates would be listed for that batch, and "gwtoolset-url-to-the-media-file", which listed the link to the image to be uploaded. Some unnecessary metadata fields, such as notes for NLS cataloguers, were scrubbed from the xml files for upload.
I test uploaded images from NLS collections on the Beta cluster approximately 3-4 times before attempting a live upload. In part this was to avoid any errors in the live upload process, but it also allowed me to experiment with the different resolutions available for upload, building a case to release images as large as 2500px .jpgs. Once the release of 2500px images was confirmed and the process was determined to be relatively error-free, I requested access to the live version of the GWToolset, once again done through the Glamtools mailing list, and began uploading the images to Wikimedia Commons itself.
Before uploading content, I created categories that would make it easier to find and network the images in Commons—in addition to creating a category for images uploaded as part of the collaboration between WMUK and the NLS, I also created categories for the collection being uploaded (i.e. Category:Weir Collection) or related to the content itself (i.e. Category:Construction of the Forth Bridge) and nested those categories where appropriate. All content was also uploaded to Category:Images from National Library of Scotland.
Incorporation and dissemination
The short timespan between the uploading of NLS content to Wikimedia Commons and the release of the 12 month report means that metrics pertaining to the incorporation and dissemination of the material throughout Wikipedia and other Wikimedia projects is still unknown. Steps to track the usage of the files in the future include the following:
- Track any increase in image use across Wikimedia projects via the BaGLAMa2 tool. The category 'Images from National Library of Scotland' has already been added for this purpose
- Track the rate of traffic directed to the NLS’s Digital Gallery from Wikipedia sites and Wikimedia Commons using Google Analytics
- Track the number of monthly views for articles on English language Wikipedia included in the category 'National Library of Scotland' using the Treeviews tool
- Assess and improve the categorisation of content relating to the National Library of Scotland both on English language Wikipedia and Wikimedia Commons where appropriate to facilitate previously mentioned methods
Information on the metrics relating to uploaded content will be made available on an ongoing basis throughout the continuation of the residency on the National Library of Scotland’s GLAM project page.
