Geotagging Web Pages and RSS Feeds

My previous
article discussed how geolocation by IP address
enables applications and Web sites to determine users' locations
automatically in order to provide specific location-based services to users and members of
an on-line community. In this article, we present various methods by
which Web sites can provide their geographic locations to static
pages and syndicated feeds, in the form of meta information or
geotags. Put another way, geolocation by IP address is
the technique a Web site uses to determine where users are located; geotagging is the technique
users employ to find out where a Web site is located.

Geotags typically locate the Web site's principle location on the
Earth. This information can include latitude and longitude information
for exact locations or simply city, region and country information for general
locations. Web services, applications and users then can query this
information to obtain directions (how to get from here to there),
locality (what's near there) or context (where was this article
written). Geotags differ from a simple address in that they usually are
encoded in metadata and are not visible as part of the Web page. Furthermore,
by following a standard, other services easily and reliably can find these
geotags. Various semantic Web projects still are solidifying geospatial
tagging standards, but several techniques already have become common
and supported. This article presents these current techniques.
Why Geotag?
Providing a geographic location is beneficial particularly for retail and
service businesses, tourist attractions and entertainment
venues. Geographic link directories, such as A2B and
Multimap, can index these services by
location and allow users to search geographically as well as by service
type. Currently, many of these services limit users in their selection of
available services. But, it would be possible to allow for more complex
queries, such as searching for "Italian restaurants within 2 miles of
downtown Arlington, Virginia". Or, when using automatic geolocation, one
could ask for "directions from my current location to the nearest theater".

Current location-based services rely on the Web site administrator
registering with an on-line index and specifying its location. Some
of these services charge a fee, and many are not used commonly, nor are
they cross-referenced. Google is now beta-testing
Google Local, a free engine that allows users
to search for location-based services using complex queries such as
the examples above. Unfortunately, Google Local probably uses a form of scraping
instead of meta information, considering that the
2002 Google
Programming Contest winner's entry, by
Daniel Egnor,
is a geographic search which: "includes
a geocoder (... to turn street addresses into latitude/longitude
coordinates), a simple indexer that looks for addresses and keywords in
documents, and a query engine to search for documents matching certain
keywords that also contain addresses within a certain distance of a
target location". Still, Google Local is an excellent example of how providing
geographic information on a Web site greatly can enhance its visibility
and usefulness to potential customers and users.

Geographic metadata also is useful for bloggers and
photographers. Traveling writers, travel writers and reviewers
can give context to their articles by supplying specific geographic
information about where they are writing from or where the business
they are reviewing is located. Photographers can provide viewers with
information necessary for better understanding the photograph by
informing them of where it was taken. Environmental services are now beginning to offer
syndication feeds for
weather and
earthquakes. By geotagging these feeds, aggregators
could sort, search and display information by region and location. As a
result, users would gain a better picture of the current events
happening in their areas.

By embedding a geographic location in the metadata of the Web site,
applications and Web-based services quickly and reliably can determine
the site's location relative to search criteria. Using metadata prevents
the confusion of an automated search bot having to determine the
location from the site's text. The rest of this article discusses the
techniques used for embedding geographic information in your Web site
or syndicated feed.
Geotagging a Web Site
For a Web site, several means of geotagging are available. My
previous article explained how to embed the site's geographic information
in its DNS entry. Other options also allow this information to be placed within a site
or each individual page. These are the older ICBM tags and the newer,
more generic geo-structure tags.

The ICBM (original acronym is Intercontinental Ballistic Missile)
tags derive from a more historical application, as described in the
AntiOnline
jargon dictionary:

(Also 'missile address') The form used to register a site with
the Usenet mapping project, back before the day of pervasive Internet,
included a blank for longitude and latitude, preferably to seconds-of-arc
accuracy. This was actually used for generating geographically-correct
maps of Usenet links on a plotter; however, it became traditional to
refer to this as one's 'ICBM address' or 'missile address', and some
people include it in their sig block with that name. (A real missile
address would include target elevation.)

ICBM tags are limited to latitude and longitude and do not include
other regional information, such as city or country. From Matt Croydon's
PostNeo, the RFC (request for
comment) of the syntax is as follows:

  

<meta name='ICBM' content="latitude, longitude" />

This tag would be included in your Web page's <head> section.

Another means of embedding geographic metadata is through geo-structure tags.
These geo-structure tags can include latitude and longitude information
as well as regional information and an extra placename. The
placename could contain the specific address of the person or business.
Or, it could be useful for providing a location that may not have a specific
point but covering a broader region, such as a city or district. The following
example is for the Museo Nacional Del Prado, in Madrid, Spain.

<meta name="geo.position" content="40.4157;-3.6947" />
<meta name="geo.region" content="ES-M">
<meta name="geo.placename" content="Paseo  del Prado">

I obtained the geo.position information using Multimap, given the
address provided on the museum's Web page. The geo.region uses the

ISO-3166-1 Country Names and Region Names specifications. For the US and Canada, this is
the abbreviation for the state or province; it varies in other countries. All
together, the code for my Web site looks like this:

<meta name="DC.title" content="High Earth Orbit" />
<meta name="ICBM" content="42.4266, -83.49307" />
<meta name="geo.position" content="42.4266;-83.49307" />
<meta name="geo.region" content="US-MI">
<meta name="geo.placename" content="Northville">

As mentioned in the introduction, Multimap provides
a location service that offers hotels located within the area of the maps you have
selected. This service is provided by the
Accommodation Search Engine
Network. The Web pages on the ASE site employ the
use of commented-out tables to hide the geographic information. The
Multimap Web service then pulls this information from each of the
hotels' Web pages or backend databases. This method is not recommended,
however, because it is not standard and is difficult to use for applications and other
Web services. An example of this block is shown below:

<!--
<a name="geo"></a><br>
<table>
  <tr class="bg2">
    <td class="head2">Geographical Information:</td>
  </tr>
  <tr class="bg1">
    <td class="text">
      <span>Accuracy:</span>Property<br>
      <span>Lat:</span>39.894337<br>     
      <span>Long:</span>-83.807657<br>
	  <span>TimeZone:</span>Eastern<br>
    </td>
  </tr>
</table>
-->

Geotagging an RSS Feed
Besides geotagging a Web site, it is possible to geotag the source of an
RSS feed as well as the individual articles. By geotagging each article,
your feed can provide entries from various locations. Then, these
entries can be displayed on a map where users can read about locations
that interest them. Alternatively, by geotagging the source of the feed, a directory
or opml file could provide feeds based on user-selected locations.

The Resource Description Framework Interest Group (RDFIG)
has published
a solidifying standard for
the geospatial vocabulary. It includes specifying
latitude, longitude and, optionally, altitude. This is similar to
the geo tags discussed above, with small syntactic differences. Furthermore,
the RDF requires you to specify an XML namespace (xmlns) for the
WGS84 geodetic reference datum.

<rdf:RDF 
        >
  <geo:Point>
    <geo:lat>55.701</geo:lat>
    <geo:long>12.552</geo:long>
    <geo:alt>52.4</geo:alt>
  </geo:Point>
</rdf:RDF>

The ICBM standard discussed above also can be used in tagging an
RSS feed. Again, an XML namespace is used to specify the keywords
of the file, and the tags are included either in the header
or within the item tags. Here is example from the
USGS Earthquake
feed of events over 2.5 on the Richter scale in the last 7 days:

<rss version="2.0" >

<item>
	<title>M 3.7, Southern Alaska</title>
	<description>January 02, 2005 03:55:52 GMT</description>
	<link>http://earthquake.usgs.gov/recenteqsww/Quakes/ak00043775.htm</link>
	<icbm:latitude>60.4780</icbm:latitude>
	<icbm:longitude>-152.4355</icbm:longitude>
	<dc:subject>3</dc:subject>
	<dc:subject>pasthour</dc:subject>
</item>

Finally, some Weblog services may prevent users from adding new tags to RSS
feeds. In this case, it is acceptable for some sites and packages to
embed the geographic information in <dc:subject> tags, as shown below:

<?xml version="1.0"?>
<rss version="2.0"  >
<channel> 

<item>
	<title>Example Title</title>
	<link>http://site.com/geo</link>
	<description>Example Description</description>
	<dc:subject>geo:lat=33.00 geo:long=-44.54</dc:subject>
</item> 

</channel> </rss>

Several Weblog packages already incorporate the ability to specify a
geographic location within an entry as well as for the entire Weblog. This
geographic information then can be included for users when reading the
Weblog through their browsers or through their own aggregators. Each
entry, when posted, is assigned either a default location or is given
a new location. The following is an example of the geotagged RSS feed
generated by Wordpress v1.2:

<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" 
	
	
	
	
	
>

<channel>
	<title>High Earth Orbit</title>
	<link>http://highearthorbit.com</link>
	<description>LinuxJournal Example</description>
	<icbm:latitude>42.4266</icbm:latitude> 
	<icbm:longitude>-83.4931</icbm:longitude>	<copyright>Copyright 2005</copyright>
	<pubDate>Wed, 05 Jan 2005 18:54:47 +0000</pubDate>
	<generator>http://wordpress.org/?v=1.2</generator>

<item>
	<title>Sample Title</title>
	<link>http://highearthorbit.com/index.php?p=121</link>
	<pubDate>Wed, 05 Jan 2005 18:54:47 +0000</pubDate>
		
	<category>Sample Category</category>		
	<guid>http://highearthorbit.com/index.php?p=121</guid>
	<description>Sample Description</description>
	<geo:Point>
		<geo:lat>42.5021</geo:lat> 
		<geo:long>-83.1454</geo:long>
	</geo:Point>
	<icbm:latitude>42.5021</icbm:latitude> 
	<icbm:longitude>-83.1454</icbm:longitude>
</item>
</channel>
</rss>

For more information on other particular Weblogs, check out the
Worldkit
Documentation.
What to Do with Geotags
Now that your Web site has been geotagged, what can you do to share
this information with users and have new users find your site? A2B
is the new incarnation of the defunct geourl.com. A2B
allows Web site administrators to register their sites. From there, users can
search for sites based on location or geographic locality to another Web site.
It may be interesting to find out what other sites and places can be found in your area.

A2B also provides a
free public
API that allows application and Web site developers to query
the A2B database of locations. The A2B query does not return the actual
location of the Web sites, however, merely their distances and
directions (compass headings) from the queried location.

To find out the latitude and longitude or city and region of a Web site,
the user can view the Web site's meta information. To illustrate
this, we have written an extension to the Firefox browser that alerts
users that geotags are available for the Web site currently being
viewed. The extension also retrieves that information without the user having to look at the
Web site's markup source.
Download
and install the extension to try it out for yourself.

Another index to check out is
WorldPress.
For RSS feeds,
MapBureau and Michael Maron's
WorldKit
Mapper have on-line mapper applications that parse out the locations from your feed and display
them on a map. It then is possible to embed a link to a map of your feed
in your Web site.

Other applications of geotags include creating a Web page of closely
related Web sites, similar to a Web ring, and display their locations on
a map of the Earth or a specific region. A restaurant review Web page,
for example, could display a map of their reviewing regions, and users could click on
locations to read reviews of the restaurants located there. Furthermore,
travelers could pull up Weblogs and travel information for the area
they will be visiting. Hopefully, larger services similar to
Google Local or Multimap will be developed that automatically will collect
and use this information to provide users with a large database of
services.
Future of Geotags
Geotags currently are not employed widely, and only a small number of
services support their use. However, many could benefit from
better geographic knowledge of Web sites and on-line data. Applications
could provide a central location to assist users in finding out about
their locations or intended travel locations. In order for this to occur,
a better standardization of geospatial metadata must be created,
utilized and supported by the Internet community. The
W3C Semantic Web is such an effort to standardize the
extension of Web data. Many groups across the globe are working
together to create enhanced definitions (see Resources). Part of these
efforts is defining a complete standard for geospatial tagging and for
supporting other location-based services. With this work, the future
of geotagging will provide better integration between the digital world
and the physical world.
Resources
GeoOnion

Open Geospatial
Consortium

Geosemantics Interest Group
(GIG)