Thursday, September 27, 2007

Volunteered Geographic Information

I'm bummed that I'll be missing the Workshop on Volunteered Geographic Information this December in Santa Barbara. VGI is at the core of the Pleiades vision.

Mike Goodchild has recently posted a paper on the workshop site entitled "Citizens as Sensors: The World of Volunteered Geography" [pdf; 1.1MB]. In the 15-page treatment, G. provides background and context; overviews a number of recent trends, key services and sites and glosses various enabling technologies.

He also introduces or glosses several underlying issues and concepts, no doubt in part laying out tracks and themes for the workshop:
  • Spatial data infrastructure patchworks (rather than comprehensive mapping strategies)
  • Humans as sensors
  • Citizen science
  • Participant populations (the who-may-volunteer vs. quality axis)
  • Early warning (e.g., of the magnitude of natural disasters)
  • Motivating factors for participants
  • Authority and assertion
  • The digital divide
G. concludes with the following reflection:

VGI has the potential to be a significant source of geographers’ understanding of the surface of the Earth. It can be timely ... it is far cheaper than any alternative, and its products are almost invariably available to all ...

It is already clear in many fields that such informal sources as blogs and VGI can act as very useful sources of military and commercial intelligence. The tools already exist to scan Web text searching for references to geographic places, and to geocode the results. Thus the most important value of VGI may lie in what it can tell about local activities in various geographic locations that go unnoticed by the world’s media, and about life at a local level. It is in that area that VGI may offer the most interesting, lasting, and compelling value to geographers.

Duck and cover?

Am I asleep, or shouldn't I have heard about this first from the Huntsville Times (links mine):
Huntsville will outfit an abandoned mine capable of holding 20,000 people. Other residents will be housed in college dorms, churches, libraries and research halls. City planners hope they can develop enough shelter space to house 300,000 people; enough space to provide every person in Huntsville and the surrounding county a safe refuge.

Maybe Lee can find out what it's all about.

Bibliographic Proximity

Shawn is trying out the Pleiades Atom+GeoRSS feeds in Yahoo Pipes, and is also thinking about geoparsing locational data from regular bibliography.

Seems to me it would be fun to throw something like Sean Gillies' Mush into the mix to get spatial correlations between the Pleiades gazetteer (if I can call it that in this context) and Shawn's geo-bib.

Then, if we could apply some version of the combined process to, say, the new acquisitions list of the Burnam Classics Library in Cincinnati, we'd have some nifty pre-processing that could speed identification of new works to cite in the Pleiades bibliography.

Mine, it's all mine

Technorati spiders, behold: Technorati Profile

evolve librum manualem futue!

Over at classics-l they're asking:
When is it ok to be incompetent?

Feeds for Pleiades data

Shawn Graham has tried pumping Pleiades data through Yahoo Pipes. That's exactly the sort of thing we want people to be able to do. In looking at his post, however, it's clear that we could be doing a better job of revealing our data interfaces to users. We've got a big rev coming up in October (we're moving to Plone 3), and we'll address this UI issue then. Meanwhile, maybe this post will help.

Shawn wanted to grab a feed and map its content. He did the obvious thing and navigated to our places section. Then he chose one of our pre-packaged subgroups: archaic places. Then he looked for a feed. Out of the box, Plone gives us a <link rel="alternate"> as well as an "RSS" button on the interface. Transiting the corresponding URL gets you an RSS feed listing titles, descriptions and other Dublin Core metadata for whatever Plone content is surfaced at that location.

As Shawn observed, pumping that list of ancient names through a presentist geocoder (like Yahoo!'s) gives you suboptimal results.

Pleiades in fact stores locations for every feature (at least when we can determine their locations). In our customization work on Plone for Pleiades, we've tacked on a couple of other interfaces that aren't as obvious to users as they should be. Anywhere Pleiades displays or lists spatial content, we also provide an Atom feed that's extended with GeoRSS tags, as well as a KML feed.

So, for those archaic places, Shawn could choose to use either of:
Both provide the coordinates, and therefore get you around the geocoding problem.

Pleiades needs to add <link rel="alternative"> tags for both our Atom and KML serializations, as well as clear UI hooks everywhere the feeds are surfaced. We do the latter in some places already, just not everywhere. See, for example:
Shawn: thanks so much for taking a hack at this! It's great to know that folks out there are interested in our data, and interested in using it in some of the same ways we've been thinking about. We'll try to meet you closer to half-way next time.