Wikimedia Engineering/Maps & Geo Team

Proposal

Throughout the last number of years the question of Maps and Wikimedia has come up multiple times but has gotten little traction due to lack of staff resourcing. This proposal focuses on building and staffing a team of engineers and operations staff members to support a product pipeline of map and geospatial features. This will address the fact that our current proposal is unsustainable and that our user experience does not reflect a modern encyclopedia when 20% of our pages already feature geo location information and 33% of edits [CITATION] but our lack of tooling hampers additional growth of high quality content.

Background

While the foundation has made small strides in learning more about maps through initiatives like the Wiki Loves Monuments App and Special:Nearby the velocity of progress has always been limited by a lack of operations and engineering support to provide a strong and supported maps foundation. For instance, the build out of infrastructure support for OpenStreetMap has been slow and haphazard not simply due to the complexity but mostly due to priorities shifting within various teams and no one having enough time to focus on maps. At the same time we've had a significant amount of community interest and contributions for maps and geo contributions. This could be greatly increased if proper software support is added for maps and geo infrastructure.

Team

To remedy this I propose building a Maps and Geo team incubated within the Mobile Department that would have 4x engineers, 1x operations staff member, 1x community liaison, and 1x product manager. These roles may live within mobile or be matrixed into mobile if they are 100% dedicated to Maps and Geo work. This engineering team would live within Mobile as they would work very closely with the Mobile Web & App teams at first in order support their initial product work but would serve as an infrastructure team that would eventually build maps and geo functional for the organization at large across handset, tablet, and desktop. The mobile team has expressed the most amount of interest and drive for this work and the department is in a good place to absorb additional growth due to its Agile practices.

2 Engineers

The main focus for a two person engineering team is to create a stable tile infrastructure and do a basic triage on existing tools

• Building and maintaing a production tile service that supports Mobile Apps, Web, Multimedia, and Community Development
  • Scale the service to work at Wikipedia traffic levels and third party use
• Take *some* action on maps and geo technical debt
  • Update geohack to function within Wikipedia rather then a third party site
    • Currently users are taken away from Wikipedia to a site that looks like nothing of what they were on
  • Update WikiMiniAtlas to be supported by internal tile service
    • Currently maps are slow and don't even load consistently
  • Update synthetic maps like File:California_county_map to be easily creatable
    • Currently they are all manually and inconsistently created
• Build and support storage for tile service
  • Current projected storage is barely enough for redundancy
• Supporting an OSM Database for tool labs developers
  • Current tools include
    • Wikipedia-World, POI Tunnel, WIWOSM, OSM-Gadget, Query-to-map, Multilingual Country-List, etc

4 Engineers

A four person team will be able to focus on the above plus they will have the capacity to develop new features and extend to other teams

• Create custom tile sets in collaboration with MapBox and Stamen design
  • Instead of the default OSM tiles we'll be able to create beautiful maps that represent our projects best
• Create a Maps namespace to support custom map creation
  • Creating navigable maps within Articles using tile rendering service
  • Create custom maps through usage of GeoJSON. No need for templates
• Create a map and location based navigation model for Wikipedia to augment search
  • Use maps and not just search to browse Wikipedia
• Work with WMF engineering teams like Mobile and Multimedia to extend Maps and Geo functionality
  • Support in app maps use and campaigns
• Collaborate with Wikimaps project to easily allow spatial map creation
  • Allow for import of 2d maps and create navigable dynamic maps
• Create workflows to add GPS coordinates for articles that don't have them
  • Surface articles that need GPS coordinates with correlate a user locations and article categories
• Evolve Special:Nearby for desktop and mobile users
• Increasing the amount of Wikidata contributions
• Increasing the amount of GEO-tagged coordinates in the Global South
  • Work with Wikipedia Zero team to increase Geo contributions in Africa, Central and South America, and Asia

Outcome

As our mobile page-views continue to climb Wikipedia will be more and more present within the physical world. Not having maps will create an inferior product that doesn’t deliver the rich educational experience that we need to be showing. Instead we should focus our efforts at staffing this initiative and pairing it across Mobile, Multimedia, and others in order to capture the wealth of contributions and use that this will have. Creating a rich Maps and Geo experience will broaden both our readership and contribution base and will modernize our encyclopedia to have an experience that will lead us in the future.

Required capacity analysis

The first things we want to try our own tile server with is our mobile apps - we can predict that the number of map views for them will correlate with current Nearby usage. Our Nearby usage appears to be the main source of GeoData usage, which is currently at ~250k requests/day.