Introduction to TileMill

From DevSummit
Jump to: navigation, search

Facilitated by Eric Gundersen, Development Seed


TileMill is one of the most exciting platforms to emerge over the past several years. Eric will explain what makes it different from other tools as well as the range of ways in which nonprofits can leverage the technology in their programmatic work


Introduction to TileMill

  • TileMill is a tool for offline mapping.
  • There is no such thing as "neutral data visualization."
  • MapBox is a company that hosts maps.
  • The space of mapping is going to get much more complex in the next 12-24 months.
  • Using data and data visualization in a narrative context is key for complex issues like climate change.
  • Certain stories should be "masked" to maintain a tight narrative.
  • Overlaying data on maps is the most traditional method for this type of narrative.
  • Treat maps as "context canvases".
  • Controlling the design is key to "selling" the map.

TileMill at Larger Scales

  • One advantage of TileMill is how well it works with big data (e.g., NPR used it for analysis of 2010 census data.)
  • It's construction is fast and light, and doesn't noticeably slow even with big sets of data. It does this partially by caching big data sets.
  • It allows historical snapshots of a specific moment in time even if the data was not input "live" at that time.
  • It also supports adding Google Chart elements

TileMill Lets You "Glue" Many Open Source Projects Together

  • It lets you construct open layers for complex and spatial work (with Drupal).
  • It works with MapBox to make it relatively easy to design and publish maps.

General Procedures and Options

  • The order of operations is "Data Collection->GIS Analysis->Map Design
  • TileMill is a downloadable application for desktop design, so it can be used without internet access.
  • Data needs to be geocoded to be useful.
  • MapBox has a Google Docs plugin.
  • There is also a CartoCSS interface built into TileMill which lets more advanced designers do coding by hand.
  • Offline maps can be stored and retrieved in the .mb file format.
  • Some of the layers you can add to the map are terrain, language, and data.
  • TileStream is recommended as an open source tile server.
  • Single tiles run on JSON which allows user interaction.
  • You can set up your map to integrate with dynamic data so it is always up-to-date. One example is an Internews map which tracks locations of violence against journalists in Afghanistan. It also lets you combine "slow" and "fast" data points.
  • Because it's open source, there are libraries of code samples which let you hack it and have complete publishing control even without a lot of coding experience.
  • You can also use a Google Docs survey form for continuously collecting data. Another tool, the Fulcrum application, works like SurveyMonkey, but for teams.
  • It can also integrate with Open Data Kit collecting data from custom mobile apps.
  • Even (or especially) when you're dealing with lots of data, avoid overloading the user interface.
  • GeoIQ (now owned by ESRI) is another great resource for geocoding and research.
  • Another great open source tool is Open Street Maps.