Sunday, May 19, 2013

October 2012 CyberSelection: GeoNames



GeoNames (www.geonames.org) looks plain, unexciting, and bare-bones, but it hides a lot of power behind its dull facade. Founded by Marc Wick, a self-employed software engineer in Switzerland, the site contains over 10 million geographical names from all over the world. GeoNames exists commercially as a web services tool for developers; you may have benefited from it when using sites such as Bing Maps, the U.S. Geological Survey, the BBC, and Nokia. Much of the data that moves below the surface in those sites is available for free at www.geonames.org, and it can be used directly by individuals in a variety of ways.

Although it has been around since at least 2006, I only recently discovered GeoNames as a partner to The European Library (www.theeuropeanlibrary.org), where it was mentioned as an authority file for geographic names. At the GeoNames site itself, I soon found out, you can enter the name of just about any city, state, region, country, lake, river, sea, mountain, or a slew of other location or geopolitical entities, and find out what and where it is. When I lost track of all the foreign place names I was reading in The Poisonwood Bible, for example, I entered Kikongo into the search box at GeoNames, clicked Search, and received a page (www.geonames.org/search.html?q=kikongo&country=) telling me that Kikongo is a "populated place" in the Congo at S 5° 50' 12'' latitude and E 15° 1' 24'' longitude. Clicking on the hyperlinked country name Congo, I got a country page (www.geonames.org/countries/CD/congo.html) listing basic facts about the Congo: country name, ISO code, FIPS code, capital, area, population, currency, languages, the names of neighboring countries, and a tiny view of its national flag. There were also a line drawing of the country and links to other sources of information (Administrative Division, Feature Statistic, Largest Cities, Highest Mountains, Other Country Names, and Wiki).

The site is not completely intuitive, and explanations are minimal. If, like me, you've forgotten a lot of what you learned about geography in a formal setting, you can explore and learn by working your way down the links under Browse the Names at the left of the homepage.

Countries

The Countries page (www.geonames.org/countries/) presents a table of over 300 countries, with columns for the country name, capital, area (square kilometers), population, and its continental home. Clicking the country name leads to a page like the one described above for the Congo, though some have more complex data. Denmark, for example, shows "dependencies" (Faroe Islands and Greenland) and Puerto Rico shows that it "depends from" the United States. GeoNames uses English for all country names, but an extremely useful feature resides below the "other languages" link next to the country name: a table of names for that country in the other languages of the world. So, for example, if you want to know how they write "United States " in Afrikaans, Arabic, or Azerbaijani, just find the right language (click on the "language" column to sort alphabetically).

In addition there are three columns for variations of the ISO 3066 country codes: the two-character alpha codes, three-character alpha, and three-digit numeric. I am familiar with many of the two-digit country codes (US for USA, CA for Canada, DE for Germany, ES for Spain, etc.), though I would have been hard pressed to name the international standard that defines the codes. A fourth column gives FIPS codes, which I had to look up; FIPS is in fact an outdated standard, withdrawn by NIST as a Federal Information Processing Standard in 2008. I'll bet there are still some computer systems that have not made the changeover, though, so I see why it would be useful for the commercial part of the site. Any of the columns on the Countries page is sortable by clicking on its header, once to put in descending order; click again for ascending.

Cities, Mountains, and ZIP Codes

The Largest Cities link from Browse on the homepage leads to an interactive Google map and chart of ten cities. You can move the map to re-center and select another part of the world, then select one of the pinpoint locations to get a pop-up window with more detail. I moved the map to South America to view cities in northern Argentina and adjoining countries. The ten-city chart shows the country, population, and "feature" of each city. I was puzzled to see that all my Argentine cities were "seat of a first-order administrative division" except for Buenos Aires, which was the "capital of a political entity" and Rosario, which was "seat of a second-order administrative division." When I clicked on the pinpoint for Rosario, the pop-up window showed  a "geotree" link. GeoTree opens up a giant hierarchical outline that shows the relationship of all divisions within the entity. Sure enough, Rosario is the seat of the Departamento de Rosario, which is one of 12 Departamentos of the Provincia de Santa Fe, which in turn is one of 24 provincias of Argentina. The first-order cities of Córdoba, Mendoza, and San Miguel de Tucumán that I had seen were all seats of provinces.

A similar map and chart structure appear when you click Highest Mountains from the homepage, as also happens for Capitals, though in Capitals you get 50 names in the chart.

The Postal Codes link goes to a page (www.geonames.org/postal-codes/) with line drawings of more than 60 countries. You can enter a place name, select a country, and receive a postal code; or enter a postal code, select the country and find out what location it refers to. Depending on the country, some of these are more refined than others.

Search

The keyword search box on the home page, where you are told to "enter a location name," such as Paris, Mount Everest, or New York, defaults to searching all countries, but there is a drop-down menu to limit your search term to any country of the world, from Afghanistan to Åland. You can then "search" for a table of terms matching your location, or "show on map" to bring up the Google map within the GeoNames site.

A link to Advanced Search (www.geonames.org/advanced-search.html?) appears directly under the main keyword search box and offers three additional options. You may select a "feature class" from a drop-down menu. Feature classes are:
country, state, region
stream, lake
parks, area
city, village
road, railway
spot, building, farm
mountain, hill, rock
undersea
forest, heath

The second advanced search limiter is by Continent: Europe, Africa, Asia, Oceania, North America, and South America, with a default of All.

The third option is a simple checkbox, allowing a "fuzzy search." Then, again, you may "search" for a text table or "show on map."

Authorities and Updating

The use of GeoNames as an authority file is interesting, since the site allows any registered user to edit and add new names using a wiki interface. For such a complex project on a worldwide scale, this makes sense, and GeoNames has "ambassadors" (www.geonames.org/team.html) in many countries of the world. No qualifications for ambassadorship are stated, but names and email addresses are given; ambassadors specifically help with knowledge of administrative divisions in their countries.

No qualifications other than registering are stated, either, for site editors, but recent modifications are cataloged (www.geonames.org/recent-changes.html) showing changes and who made them. A user manual (www.geonames.org/manual.html) gives directions on how to update and a user forum (http://forum.geonames.org/) shows conversations from people who seem to know a lot about geography and programming.

Flattening the World

Capturing millions of locations all over the Earth and making them searchable by name and viewable on maps is ambitious. Occasionally, especially for less populated parts of the world, you will come across a notation in GeoNames that says that insufficient map detail is available. Mistakes are also made in identifying places or their exact location (there are two entries for the town I live in, for example). But GeoNames is a first-stop resource for looking up any geographic name--great for identifying and disambiguating places (How many Medfords are there in the US, anyway?). It is also an interesting example of a successful commercial venture that still offers data for free.

Susanne Bjørner provides editorial services to publishers, librarians, authors, and researchers from a base near N 38° 3' 42" and  W 0° 47' 37". Contact her at bjorner@earthlink.net or www.bjorner.info.




No comments:

Post a Comment