The Web APIs Hub invites new and experienced developers to try Wikimedia web APIs in a friendly environment.

Problem statement

Wikimedia has made its APIs freely available, but for external developers it is difficult to figure out what they can do with our APIs, and how. Our web APIs and their documentation are currently scattered across several projects, with no central location for people to learn about and build upon them. This is the problem the Web APIs hub wants to solve.

The main target audience for this hub are developers not developing, or even running, the MediaWiki software. We want to diminish the risk of confusing them by mixing the content relevant to them with the rest of mediawiki.org. We plan to address this risk by morphing the API: namespace into the Web APIs hub, controlling and contextualizing links to other part of mediawiki.org or other Wikimedia sites. We also want to try offering a different look & feel in this namespace.

See also the ongoing list of requirements (under discussion).

Get involved

Check the Phabricator board where we are tracking all the tasks: #Web-APIs-Hub. Our goals for October-December 2015 are to:

  • Map current use of Wikimedia web APIs (phab:T114017) to guide where to focus documentation effort. In addition to this analytics work, we welcome your feedback on what you find difficult or confusing.
  • Write more articles (anyone can help, see /Contributing)

Provide feedback on this page's talk page

Structure

The hub is divided into three sections:

  • Inspire: where cool projects are showcased; links to more detailed project description pages
  • Explore: where users can play with the API(s) in functional sandboxes
  • /Build: where all the static official documentation and pointers to data sets reside.

Early wireframes:

Metrics

phab:T102079 "Metrics about the use of the Wikimedia web APIs"

We also have stats related to engagement with www.mediawiki.org. We want to filter these down to the API: (and possibly Manual:) namespace.

Requirements

Personas

  • Ankita, mobile developer willing to use Wikimedia data to enhance mobile apps.
  • Alberto, data scientist employed at an organization gathering and releasing data that could be synced with Wikimedia's.
  • Reetta, cultural activist working on mass-upload activities for public institutions.
  • Yanhui, academic researcher needing a massive data set to sustain his thesis.

These represent the users we want to make happy first (phab:T92941)

Inspire

  • A curated collection of projects created using data from Wikimedia projects.
  • Means of submitting a project to be considered for showcase.

Build

  • Documentation at different technical levels.
  • An organized list of APIs and datasets available.

Community

  • Way of requesting new datasets and/or APIs.
  • Related news, features, and announcements.
  • A way for the community to get in touch with WMF folks for help on projects.

Technical infrastructure

  • Possibility to customize the UI in order to provide a pleasant user experience.
  • Possibility to create and edit content comfortably, including the homepage, tutorials, project showcases, code examples and screenshots.
  • Possibility to export API documentation from source code in repositories.
  • Possibility to integrate API sandboxes.
  • Possibility to search all the content at once.
  • Possibility to translate manual documentation (not the strict API docs)

Internal resources for developers

The following is a list of resources housed across multiple locations to help developers build using Wikimedia data. Feel free to add to the list below.

Documentation

API Sandbox

  Done {{ApiEx }} provides a link to "Try in sandbox", defaulting to enwiki. Also featured in devhub section of landing page.

Existing APIs

Other internal(ish) projects

External examples

Developer hubs & API documentation

Best in Class human readable API documentation

Other examples:

Programmable Web

Twitter

Google

Facebook

Evernote

LinkedIn

GitHub

Amazon

Dropbox

Instagram

Stripe

Paypal

Tumblr

Last.fm

Twillio

Vimeo

Pinterest

The Guardian

New York Times

Mendeley

Flickr

Sandbox Examples

Data Sets