ARCHIATER Tutorial Uploading Datasets to Wikidata with OpenRefine

In this video we will show you how to upload a large data set from any kind of spreadsheet on Wikidata using OpenRefine. We will show you an example with an Excel table. Links: Create your project: Go to Wikidata:Tools/OpenRefine and open Paws. Once at PAWS: A Web Shell, sign in with your Wiki account. Then you will find yourself in JupyterLab and here you are able to open OpenRefine and create your own projects. For security reasons, you can not use your main Wikidata account authorization with OpenRefine. You have to generate a new and special password to access your account through software. More information and links a the Overview of Wikibase support. The project: ERC ARCHIATER - Heritage of Disease: The Art and Architectures of Early Modern Hospitals in European Cities. ARCHIATER - Heritage of Disease is a five-year project (2024-2029) funded by the European Research Council and based at the LMU, Munich (grant agreement No. 10109/906). It will analyse the spaces, forms, themes, and designs of hospital art and architecture, and reconstruct the transitions and changing values associated with hospital materiality and aims to investigate the role of artwork in hospitals and explore how visible artifacts can contribute to the well-being of societies facing invisible threats. Further reading: Baumeister, L. / Franceschini, C. / Kristen, M. / Leipold, F. / Limmer, I. / Siebert, M.: “Mapping Hospital Heritage with WikiFAIR: A Proposal for Research Data Management”, in: Zeitschrift für digitale Geisteswissenschaften [forthcoming]. Video produced (without AI) by: Lily Baumeister and Isabella Limmer ––– Abbreviated Transcript of Tutorial Video: From an Excel table to a Wikidata entry in 10 minutes. Learn how to easily transform a spreadsheet into semantic data — with the help of OpenRefine. Go to OpenRefine via Wikidata:Tools/OpenRefine, open Paws and click at PAWS: A Web Shell. A new window opens, and you have to sign in with your wiki account and get to JupyterLab. Upload your dataset as a spreadsheet here. Check it in the preview and adjust your table, e.g. make sure that the first row is the column headers by clicking parse 1 line(s), then create your project. If necessary, adjust your table even more by spliting the column or correcting small editing errors via edit cell and common transforms and trim leading and trailing white space. The most important step is to connect the data to Wikidata entries, values or qualifiers – this is called reconciling. For that we go to a column, click reconcile and start reconciling; choose Wikidata. Click next and choose a property. Check if the reconciled item is the correct one. If not, click search for match or create new item. However, before creating a new item be careful to check again if it may already be on Wikidata with a diJerent name to make sure that you don’t create a second dataset and lose information by not connecting the two entries. Repeat the process for all properties (except open text fields in qualifiers like page(s) or identified in image by). To simplify and speed up the process the table should be written with the possibilities of Wikidata in mind: no long sentences or explanations; dates/numbers only; key words and Wikidata statements; Q-IDs. When everything is matched, create the structure for how and where your data should appear on Wikidata, the so-called schema. Click at Edit Wikibase schema and add an item. Drag the first column (with all your research objects) in. Then built your schema: add statements by searching them manually and dragging in the corresponding columns consisting of properties and values. Check the Preview and possible issues. Then upload your prepared dataset to Wikdata. For this last step, you need a special password for Uploads [generate the password here]. Now you can search, visualize and work with your data directly in Wikidata or add even more information by uploading another dataset via OpenRefine.