My feeling is that we'd need to separate out the data view by projects (wikipedias, wikisources, commons, Wikiversity etc.) because trying to put everything in one view will make it more confusing than convenient for the user since the different datasets will have different headers. And splitting it out makes it possible to tailor the download buttons to the project so we don't end up giving the user one gigantic csv or one huge wikitable with everything crammed in it.
I'm trying to also think about how abstract-able this would be so we can build it more easily. Though nothing is straightforward with this tool. :)
Here's a mock that I thought might work:
I have a few questions, mainly for @MusikAnimal -
- What do you think? Other ideas?
- If the tab idea seems good, do you think we can fetch all data at once so the tab switching can be seamless? We're okay with it taking long (~2 minutes) to generate that data.
Good? Bad? Ugly? Let's brainstorm!