Wikipedia:GLAM/Museum of New Zealand Te Papa Tongarewa/The whole GLAM package/Make Wikidata edits
 | How to use this manual · Set up your project · Select your topic · Select images · Make Wikidata edits · Upload images · Create and edit articles · Review and promote · Project checklist |
Before you re-export your record data, you want to get Wikidata sorted. Now that you’ve decided on your records and images you can build a list of items you need to update and get QIDs for.
Before making bulk Wikidata edits, it’s good to know how to do them manually too. This resource takes you through the process using scientific collectors as an example.
Guidelines for using Wikidata to mobilise information about people
What to look at[edit]
Go back and see what types of items you wanted to use. This might be creators or collectors, scientific names, locations, art movements, or kinds of objects depicted.
When looking over your data you might decide you want to focus on different types of Wikidata items. Talk it over as a team and think about any impact that will have on the time or effort it takes.
There’s probably a WikiProject that covers your topic, such as WikiProject Taxonomy, so look around. You’ll find guidance on how to pick important fields, use these Wikidata items in articles, and connect different items together.
Export data and check for Wikidata items[edit]
Checklist task: Review supporting data
Before you load images, you want to make sure their supporting information is in Wikidata. You do this by exporting the relevant information, putting it into a new OpenRefine project, and reconciling that with Wikidata to see if you can find matches. This will tell you if you need to create new items or add extra data to the ones that already exist.
Make lists of the people, scientific names, and other associated records that are relevant to your selected records and images.
Export a spreadsheet for each type of associated record. For example, you can do one for all the creators, and another for the organisations. If you already have Wikidata QIDs for them, add those as well.
 | Tip If you’re putting up data about living people, make sure you’re not going to breach their privacy. Check out your local privacy laws and organisational policies. |
Make a new OpenRefine project for each spreadsheet, and start reconciling to find existing Wikidata matches.
How to use OpenRefine with Wikidata
Detailed guide to reconciliation
When you’ve made matches, it’s a good idea to add them to your collections database so you can use that connection again.
_-_The_Noun_Project.svg){kind=small} | Example
Te Papa adds Wikidata QIDs to records in EMu, our collection management system, when we’re sure about a match. For example, we can put a person’s full Wikidata URL in a field called “Web associations”, and just the QID in another called “Identifiers”. Then the QID is available for future data exports (or just a staff member looking up the record), and the link shows on that person’s page on our Collections Online website.
 |
| Tip This is another good opportunity to review and maybe update the data in your own system. |
You should now know which records have a match in Wikidata and which ones are going to need new items.
Prepare and load your Wikidata edits[edit]
Checklist task: Prepare and load Wikidata edits
To load your updates and new Wikidata items, you need a schema – this is a conceptual data structure that you will set up within OpenRefine to put each bit of information in the right place.
Your schema doesn’t need to cover every bit of information you may have in your system. Instead, you want to look at existing Wikidata items and data models and pick out just what’s going to be useful.
Remember, information in Wikidata is structured as statements, so your schema will need to be structured that way too. You need to pick a Property that helps describe some aspect of the item, and decide on a field that will fill in the appropriate value.
| Tip While most Wikidata properties need to point to items, there are some that should just be plain text. For example, an image might depict (Property P180) Myosotis glabrescens (Item Q17416556), but the taxon name (Property P225) will just be “Myosotis glabrescens L. B. Moore”. |
Look through Wikidata’s WikiProjects to find guidance about how to model different kinds of items.
| Example For people, Te Papa wants to cover the essentials and enough specificity that it’s clear who the item refers to, particularly by including the New Zealand context. For species names, we want the authorship details, genus and family names, and common names in both English and Te Reo Māori if available. |
Re-export your data with just the fields you need for your edits and new items – the most important thing is to include the QIDs you found so you can easily match them up again.
Do one more check of your data, add your schema within the project, and make your edits!
When done, review your changes and don’t forget to add the QIDs for new items into your collections database.
Next up: Load images and collections metadata