Comments on: More usefully cool stuff from Stamen

By: When does afternoon begin?

When does afternoon begin? — Sun, 21 Jun 2009 16:04:27 +0000

[...] about this same principle in another context. The new version of Oakland Crimespotting, which I raved about, segments incidents using this [...]

By: When does afternoon begin? | Tech-monkey.info Blogs

When does afternoon begin? | Tech-monkey.info Blogs — Mon, 15 Jun 2009 21:11:31 +0000

[...] about this same principle in another context. The new version of Oakland Crimespotting, which I raved about, segments incidents using this vocabulary: light, dark, commute, nightlife, day, night, swing [...]

By: When does afternoon begin? « Jon Udell

When does afternoon begin? « Jon Udell — Mon, 15 Jun 2009 20:45:32 +0000

[...] about this same principle in another context. The new version of Oakland Crimespotting, which I raved about, segments incidents using this [...]

By: Michal Migurski

Michal Migurski — Sun, 07 Jun 2009 01:08:45 +0000

There’s a lot more information on the Flickr / WOE shapes here: http://code.flickr.com/blog/2008/10/30/the-shape-of-alpha/

“This is the important part: Many, if not most, of these shapes will look a little weird. Possibly even ‘wrong’. This is both okay and to be expected for a few reasons, at least.”

By: Jon Udell

Jon Udell — Sat, 06 Jun 2009 13:45:58 +0000

OK, having read up on WOE (Where On Earth) IDs and their use in Flickr, I think I see what’s happening here. The idea is not to naturalize the query for place, but to naturalize the names under the crosshairs as you pan around after having searched for a conventionally-named place.

So when positioned over Long Island, successive zooms yield:

New York
Suffolk County
Sterling Forest

And from there, panning yields adjacent neighborhoods:

Upper Greenwood Lake
Vernon

But..searching for Vernon takes you to a different place, because Vernon the town isn’t the same WOE ID as Vernon the neighborhood next to the Upper Greenwood Lake neighborhood, and search isn’t WOE-aware.

“users’ geotagged photos are used to determine the outlines of variously-thought-about geographic areas”

So that’s how WOE shapefiles are defined? By clustering the tagnames in geotagged photos and associating them to irregular areas?

Very cool. Thanks for the heads-up on this!

And you’re right. This does appear to be an important step in the direction of the idea I sketched.

By: Eric Rodenbeck

Eric Rodenbeck — Sat, 06 Jun 2009 01:02:12 +0000

There’s a technical description of the project here:
http://www.aaronland.info/weblog/2009/06/01/bubblegum/#iamhere

and a little more about Clustr here:
http://code.flickr.com/blog/2009/04/07/the-only-question-left-is/

The basic idea is that it shows, based on zoom, how people tend to name an area…

By: Jon Udell

Jon Udell — Fri, 05 Jun 2009 22:22:41 +0000

Seems interesting. Unclear what it’s intended to do though. Tried these queries:

Keene NH -> map centered on Keene

New Hampshire -> map centered on Concord (the capitol)

southwestern New Hampshire -> map of San Diego (huh?)

By: Eric Rodenbeck

Eric Rodenbeck — Fri, 05 Jun 2009 21:28:02 +0000

The vagary of ‘state-level’ could perhaps be usefully addressed by the work that our friend Aaron Cope is doing at flickr, where users’ geotagged photos are used to determine the outlines of variously-thought-about geographic areas:
http://aaronland.info/iamhere/

“My neighborhood” means a very different thing if you’re in London or the middle of Montana.

By: Jon Udell

Jon Udell — Fri, 05 Jun 2009 16:59:11 +0000

> It might be fudgable by taking
> proximity-of-the-focus-point-to-the-
> nearest-state-border into consideration
> as a weighting factor for zoom level.

That’s exactly the kind of thing I had in mind. It ain’t gonna be perfect. But it might be really useful. And the process of thinking through what the vocabularies would need to be would, in itself, be incredibly instructive.

By: Gumnos

Gumnos — Fri, 05 Jun 2009 16:19:51 +0000

Ambiguity of “state-level” zoom requires the user to express intention. For many common cases, it works just fine. If I’m looking at Harrisburg, PA, then “state-level zoom” likely means “show me all of PA”.

However, if you’re looking at, say, Texarkana, what does “state-level zoom” mean? Enough zoom to see all of TX? Enough zoom to see all of Arkansas? All of Louisiana? All of all three states? All of most of the three states? If you pan to neighboring smaller/larger states, does the zoom level change on you?

It might be fudgable by taking proximity-of-the-focus-point-to-the-nearest-state-border into consideration as a weighting factor for zoom level.

Anyways, just thinking out loud…

-tim

By: Michal Migurski

Michal Migurski — Fri, 05 Jun 2009 15:23:30 +0000

Hi Jon, thanks for the kind words! There’s a long and interesting follow-on conversation about this launch that specifically touches on the local utility of these technical features on the standout local blog A Better Oakland: http://www.abetteroakland.com/new-and-improved-crimespotting/2009-06-03

And I’m glad you’re enjoying the pie. =)