Wikipedia:GLAM/Museum of New Zealand Te Papa Tongarewa/The whole GLAM package/Select your topic

Now that you know what matters to you and your organisation, you can decide what your project will be about. From there, look around what you have available and what’s already on Wiki platforms to figure out which contributions are the best use of your effort.

Effective and useful contributions
Checklist task: Set topic criteria

GLAM organisations could work on just about any part of the Wiki ecosystem – and too much choice can make it hard to even start. If you think about what’s important, useful, and achievable you can come up with the single topic that your project will achieve.

Sources of information
Checklist task: Identify source material

Think about what sources you’ll be using when you create articles or enrich the data in your collections database. Depending on your topic you might want to focus more on academic or popular sources.

Wikipedia falls into systemic biases in the topics that get covered and the references it cites, which projects like Black Lunch Table and Women in Red seek to address. As a GLAM, you’ll have a depth of expertise and knowledge to use sources from people who might often be marginalised.

That said, take care when seeking to incorporate Indigenous knowledge (such as mātauranga Māori, Māori traditional knowledge). The Wiki ecosystem assumes that all knowledge can and should be openly and freely shared at all times. But many Indigenous peoples know that some must be held close, due to its sacred or dangerous nature, or because sharing has previously harmed them while only benefitting a coloniser.

A project team can navigate these waters more easily if it includes people from the relevant group, or already has close relationships with them. The team should also think about how the project actively benefits the community, not just how it avoids harm.

In Aotearoa New Zealand, the Kaupapa Māori research and evaluation framework operates this way – by Māori, with Māori, and for Māori – prioritising relationships built on fundamentally Māori principles.

Create your plan for what you’re going to create or improve
Checklist task: Create plan for contributions

Stop scope creep by deciding now which articles you’re going to work on. Make a list of the articles (existing or not) that are relevant to your topic, including any broader topical pages you want to enhance – and then do your best to stick to it.

If your initial goal is to create or edit Wikipedia articles about a particular topic (such as a plant or animal species), also look for articles that summarise that topic like Spiders of New Zealand. Check the articles about people, places, or concepts that are related to your topic.

Articles that are already in great shape might benefit from a new useful fact, section, or image – for example, an article about a person that describes their scientific collecting activities. An image of a specimen they collected perfectly illustrates that section!

Put a boundary around the kinds of Wikidata items you’re going to create too. Instead of trying to find ways to share every bit of information you hold, pick a few kinds that are crucial to describing your images well, such as related people, places, or subjects (including scientific names).

You’ll probably still end up making other edits (for example, a lot of non-Western personal names and many relevant scholarly research articles still don’t have Wikidata QIDs), but don’t stress yourself if you don’t think you’re being comprehensive.

Identify data and image source material
Checklist task: Review available records

From this point in the project, you’re going to be doing several loops of exporting data from your catalogue and reviewing it in OpenRefine] for selections, updates, and refinement.

Review what you have available
Start by creating a spreadsheet of all object records that match your topic. Include as much data as you can, for example:
 * Images (both how many and how large)
 * People involved (creators, scholars, collectors, identifiers, taxonomic authors)
 * Categories or other tags (depicts, related to, scientific name)
 * Record metadata (when the record was updated, if the data is public)
 * Wikidata QIDs if you already have them
 * Anything else you might want to filter by (is the species endemic, relevant locations)



Load your spreadsheet into OpenRefine as a new project.

Starting a project in OpenRefine

By looking over all the data together, you can start filtering out things you don’t want to include, like:
 * Images that are too small
 * Records that have important data missing
 * Records that have restrictions on location or personal information
 * Redundant images

Realising you need to do more work
Don’t be surprised if this review shows you a bunch of things you want to fix, from misspellings to misidentifications to information that’s just plain missing.

Stay on track by prioritising these fixes:
 * Urgent: these have to be fixed or the project can’t continue
 * Worth doing: these are probably going to end up on a Wiki site, so it’s a good idea to get them fixed before publishing
 * Worth doing, but later: these aren’t going anywhere fast, they can go into your regular stack of work

Start at the top and focus on the first two groups.

You may also find you want more or higher-quality images. Maybe there’s images for 30 of the articles you want to do but nothing for the other 10, or they were digitised when a 640 x 480 image was high tech.

Since you’ve thought about how this project connects to your organisation’s audience and strategy, this should help you make the case for prioritising some extra digitisation.

Confirm your record selection
Remove any records you’ve decided against using from your OpenRefine project. The next time you export your data you can use it to produce a cut-down dataset, which means there will be less to sort through.

When you’ve made any needed changes to your data, export it again, get it back into OpenRefine and cut out any other records you don’t want to include. Go back to your goals and workload expectations to help you confirm your final selection.

Next up: Select images