User:CKoerner (WMF)/Work/Lessons learned from maps conversation

At the beginning of May 2016 the Maps team of the Discovery department reached out to Wikipedia communities to discuss the feature of interactive maps. Now, roughly a month later, the Maps team would like to reflect on the conversations. While the conversation will continue regarding interactive maps on Wikipedia this acts as a summary of our findings and describes the engineering team's next steps.

Findings

When we reached out to the Wikipedia communities to discuss interactive maps we asked 3 specific questions.

1. What types of articles will use interactive maps?
2. How do these articles differ in their requirements?
3. Are there any classes of articles whose map styling requirement is fundamentally in conflict with other article classes, thus requiring multiple styles?

The answers we heard from the community were diverse and nuanced. A general summary of the answers to these questions are as follows.

What types of articles will use interactive maps?

Articles that represent specific locations (as apposed to general regions), both historical and contemporary, were the leading use cases for interactive maps. Wikiloves monuments, accidents, and disasters.

Other ideas included using a map to show the distribution of items across an area. Examples were World Heritage sites, animal populations, local buildings within a city (via polygon shapes), and points of interest during historical events.

How do these articles differ in their requirements?

There were no direct responses to this question. It can be assumed that the next question, and other discussions presented, broadly cover the concerns between the differences in maps - more so than the differences between articles.

Are there any classes of articles whose map styling requirement is fundamentally in conflict with other article classes, thus requiring multiple styles?

There was a small discussion around the differences between map types (political versus say relief maps). One concern is that the styles will need to be considered at the local level. A few examples shared show the complexity of geographic regions within a single country!

Additional notes

The maps team feels like they have a better understanding of what is needed - both from the user-facing and technological requirements. Some additional points of consideration that were discussed include:

• Structured data storage - paths, polygons, reusing existing maps and the data they contain.
• UI elements - translation and usability are of concern. There was also discussion off-wiki to make sure these controls are usable on mobile devices.
• Historical maps - being able to overlay and manipulate static map layers would be a very welcomed feature (see GLAM interest in the Surprises section)
• Static maps - both in their continued role and use, but also in the ability to generate them from the dynamic map tiles.
• Map styles - after the conversation was running for a few weeks the question as to what sort of styles should the map tiles appropriate for use on Wikipedia was proposed. The only feedback given was in support of using the WikiProject:Maps conventions - at least to begin with.

Surprises

• Feedback on map styles and layouts was a smaller percentage of the conversations than anticipated. The team will move forward with focusing (in the first iteration) on providing a 'standard' Wikipedia style for map tiles that follows the existing static map conventions. This allows the engineers to focus on features for interactive maps.

• GLAM interest - At first blush many of the questions on the conversation talk page seem to be unrelated. After talking to Alex Stinson, it became clear that a group of GLAM-interested editors were already thinking about how maps could be used to further enhance cultural heritage projects. This brought up conversations around topics like historical maps and reuse of content external to Wikimedia projects.
• When targeting our conversation we focused on Wikipedia editors interested in maps. What we did not anticipate was the interest from folks within the OpenStreetMap community coming to chat with us. There were discussions on their mailing list and they even put together "Welcome to Wikipedia users" on the OSM wiki for Wikipedians new to the OSM project.
• The type of map projection was a concern that wasn't anticipated. In cases of extreme polar coordinates (North and South Poles) the current projection type of OSM maps does not translate well. In which case a static map (or an alternative projection of OSM data) would better suit the mapping needs.

What's next?

Technical work continues. Feedback and ideas continue to be welcome and incredibly helpful. Here's a few things the Maps team will be doing in the near term.

• Setting up new servers to host map data.
  • Automating tile replication and changes (currently somewhat manual).
• Migrate existing maps functionality to the new service and improve visual aspects of it, such as geohack map.
• Hopefully approach the historical maps and adding of layers soon.