Collaboratively assessing interactive maps from government.
Want to help? Make suggestions and edits to the scoring criteria, via a pull request or issue. Thanks!
We identify 21 (+/-) characteristics of an online map that we care about, and set up a Google Form. Then we invite everyone to rate maps as they come out, and publish the aggregate ratings.
A score out of 21 is easy to talk about. And as cities make improvements, we can re-score their maps. Could produce a nice end of year report card, is highly portable to other cities, etc.
The score is a neat way to promote the principles of good engagement maps -- in order to do the scoring, you have to be familiar with all the best practices that the scoring rewards.
We could take a page from the US City Open Data Census from the Open Knowledge Foundation in presenting our comparisons of maps.
Imagine each of these is worth 1 point.
Not every map will be able to score 100% -- there might be desirable characteristics for a map that aren't possible given the data/other issues. We could make a note of attributes this map isn't even competing on (for example, a partially successful map could score 10/21 while a quite successful map could score 10/12.)
- underlying data is accessible for bulk download 1 pt
- data is accessible in a non-proprietary data format (csv/geojson, not ArcGIS GDB or Shapefile) 1 pt
- data powering the map is directly pulled from an open data site via API (ideal) 1 pt
- a complete history of all data that ever appeared on the map is kept (even if some data eventually disappears, it is still accessible in the original data source.) 1 pt
- conforms to sticky-map standards (click-and-drag pans, double-click zooms, scroll wheel zooms in/out) 1 pt
- pinch & zoom works on mobile 1 pt
- can functionally jump to addresses or regions via search bar 1 pt
- provides "deep links" so that map views are shareable from the URL bar 1 pt
- comments or other input can be provided to report bad data, and responses are tracked & accountable 1 pt
- it is possible to filter any layer with time data by time 1 pt
- available at different geographic roundups/aggregations (e.g. district) 1 pt
- normalized by area in meaningful fashion, taking into account possible statistical hiccups (like a park district with no population) 1 pt
- legend is labeled 1 pt
- available as the original location dots (e.g. geocoded location of 311 reports) 1 pt
- the time of individual dots/locations is reported when available 1 pt
- available in all (?) significant local language groups 1 pt
- fall-back for screen readers (?) 1 pt
- works in all modern browsers without a plugin (no Flash or Silverlight) 1 pt
- works on both iOS and Android, cell & tablet 1 pt
- works on all desktop browsers (Firefox, Chrome, Safari, latest IE) 1 pt
- color-blind friendly (reds/greens) 1 pt
Mapping Best Practices - slides from a presentation at NICAR by Dave Cole, John Keefe, Matt Stiles
When Maps Shouldn’t Be Maps - blog post by Matthew Ericson
... examples of best practices for open map data?