Hey folks,

As many of you know, I’m working on a new version of and I want some input. I’m trying to group our members in metro areas (on certain pages—not search results) for SEO reasons, and I’m writing a simple script that takes raw metro data and converts it into a human-readable string. Can you guys take a quick look at this list and tell me if anything looks amiss?

Please note that certain areas will show under multiple states if they’re on a border. Also note that some smaller areas won’t necessarily be represented by this list—I’m working with metropolitan areas right now.

List clearly flawed. Will repost when I sort out the details.


Don’t you like us folks in Arizona?


  1. Atlanta, Sandy Springs and Gainesville

  2. Columbus, Auburn and Opelika

  3. Birmingham, Hoover and Cullman

  4. Montgomery and Alexander City

  5. Huntsville and Decatur

  6. Dothan, Enterprise and Ozark

  7. Mobile, Daphne and Fairhope

  8. Little Rock, North Little Rock and Pine Bluff

  9. Jonesboro and Paragould

  10. Sacramento, Arden, Arcade and Yuba City

  11. Los Angeles, Long Beach and Riverside

  12. Fresno and Madera
    *]San Jose, San Francisco and Oakland

You left out the entire north, North central, northeastern end and NW panhandle of Florida.


  1. Jacksonville, Orange Park, Ponte Vedra, Middleburg, St Augustine
  2. Lake City, Live Oak,
  3. Tallahassee, Panama City, Apalatchicola, Quincy, Montcello
  4. Pensacola, Milton, Pace, Gulf Breeze, Fort Walton Beach, Destin

Hey Doug,

Don’t complain…at least Chris didn’t miss you’re entire state–::)))

Texas…where’s San Antonio, Austin, Temple/Waco, El Paso, etc, etc?

lol, no but dang near. He missed the parts where all the smart people live.:wink: Down on the south end is where the folks live who can’t punch a pin through a ballot card properly. (Wait for the fur to fly on this one).

Good for the central part of Western Washington only. Bellingham in the northwest, Vancouver in the southwest, Spokane in the east, and Pasco / Tri-Cities in the southeast.

Hmm… I’m working from zip code statistical data. I guess it’s back to the drawing board.

[second list removed]

I removed the list because the system has changed a bunch since I posted this. I’ll post an update in a week or so.

  1. Phoenix, Mesa and Scottsdale
  2. Payson
  3. Tucson
  4. Prescott
  5. Yuma
  6. Lake Havasu City and Kingman
  7. Safford
  8. Sierra Vista and Douglas
  9. Nogales
  10. Show Low
    NEED Chandler Chris—:))))

Much better but now it is probably too detailed, at least for Texas…67?? CSA/CBSA…what’s that? Wouldn’t MSA’s make more sense?

Chris why not simply go by zipcode and include everywhere instead of just metro areas?


Some of what you are showing at least around Pensacola are neighborhoods not actual towns (Brent and Ferry Pass) and missing is other real towns around Pensacola. Not trying to pee on your parade but your initial request for input was probably the way to go because you will get real information from the folks who live there. May take a little longer but will be more accurate in the long run.

That’s why I was hoping I could use the CSA data, which is based on much larger groups, but clearly leaves a lot of things out.

MSA’s have been replaced with CBSA’s (starting in 2003, I think).

Search engines are very particular about how they handle links. If you have more than 50-75 links on a given page it’s going to harder to convince a search engine spider to follow the links lower down on that page. Thus, splitting our membership up by State and then Metro will lead to a search engine friendly hierarchy. If we just split by zip code there would be thousands of links per state. By using metro areas, we’ll have a little over 50 links on the front page (US States and Canadian Provinces), and then under 50 links on the next page (links to each metropolitan area), and finally links to each member on the 3rd page. This keeps the hierarchy under 4 levels deep (good) and keeps the links per page under 75 (also good).

In the end, it won’t matter for anyone who’s searching on the site, because if someone searches for 19147 it will skip all that and take them to the page for Philadelphia, PA. It’s mostly for search engine listing.

They’re not supposed to be towns—I’m trying to break the country up into major metropolitan areas. So neighborhoods are good.

Right now I’m trying to find a way to automate the process so that as groupings change and zip codes come and go we can just download new raw data and recreate the grouping database. If it looks like that won’t work, then we might have to rely on a more hands-on approach. We’ll see…


Just an idea. I would post this message over at! I’m sure your aware but if not a host of international datebase coders exist on that site and I’m sure someone has a idea on how to do what your looking for.


For Georgia you’ve got

Which really only gets the northwest side of Atlanta and probably the city of Atlanta.

Would it work better if you had cities that encompass the metro area?

i.e. Atlanta, Douglasville, Sandy Springs, Lithonia and Forest Park - This gets the cities on the North, South, East and West.

There are a few major ways to group locations. The most recent is the Core Based Statistical Area (CBSA) method. It’s based both on how different areas share a statistical core in terms of population, and also in terms of interaction. For example, people who live in the “Philadelphia, Camden and Wilmington” CBSA are more likely to a) live near each other and b) interact with each other. There’s no reason to re-invent the wheel here—many very talented statisticians worked hard to make CBSA data as accurate as possible, and I doubt anything we could contract would be nearly as accurate.

My main concern here is to make sure we represent this statistical data as accurately and user-friendly as possible (“Philadelphia, Camden and Wilmington” is definitely easier for an end-user to deal with than “Philadelphia-Camden-Wilmington, PA-NJ-DE-MD”). We’ll probably then build in a way to tweak the system based on user input.

As I mentioned above, I think this makes the most sense. We’ll start with CBSA data and then tweak the data manually where it needs it.

The only other option I’m looking into right now is using Neilsen DMA data, but I don’t think that’s quite right for us.