Notifications
Clear all

geographic centroid/centre

13 Posts
8 Users
0 Reactions
4 Views
 BigE
(@bige)
Posts: 2694
Registered
Topic starter
 

I have a data base with over 42k US zipcodes contain lat/lon on the "geographic centroid" as they claim. Anyone know how acurate that might be? The data came from the US Census of 2000.

I have used the info to calculate the distance in miles between zip codes and am wondering how acurate I might be able to claim.

While I'm at it.... can I get a definitive radii of the Earth in miles. So far I have about 6 different ones. I've experimented with them all. One gives a result claiming to be within 0.0001 (ten-thousandth) of a mile. The others are within 0.2 miles.

Thanks.
E.

 
Posted : November 30, 2010 2:31 pm
(@dougie)
Posts: 7889
Registered
 

Earth's mean radius is (6371.2 km) or 3958.9 miles

or did you mean circumference? (pi r*r)

 
Posted : November 30, 2010 3:25 pm
(@mightymoe)
Posts: 9920
Registered
 

The radius of the Earth depends on the model you use.
There is no real radius for the Earth because it is has an irrregular shape.
So models have been mathematically constructed to try and model the surface.
So like so many questions connected to surveying the answer is: It depends.
20,906,000 is what I used to always use for curvature and refraction calcs but....

 
Posted : November 30, 2010 3:30 pm
(@holy-cow)
Posts: 25292
 

Interesting problem.

I have no idea how they came up with a so-called centroid. I'd bet they used a quickie shot on each main post office for a specific zip code. They would be a "get you in the ball park" type of location to compare to another similar location. Comparing the distance from 10111 to 90111 would be fairly accurate. Comparing the distance from 66771 to 66772 would be far rougher, percentagewise. In fact, they might acually touch for some stretch.

As for the radii of Earth, forget it. There isn't any single correct answer no matter how you try to make one fit. What's the radii of a cantaloupe?

 
Posted : November 30, 2010 3:33 pm
(@a-harris)
Posts: 8761
 

The data would probably be keyed to post office location.

 
Posted : November 30, 2010 3:52 pm
(@kris-morgan)
Posts: 3876
 

20,906,000/5280.

 
Posted : November 30, 2010 4:33 pm
(@mightymoe)
Posts: 9920
Registered
 

There are maps of the zip code areas. The areas can be quite complex. To see what they are using as a centroid I'd pick a couple of the simple ones and do a quick centroid plot for it and see how it checks to their data.

 
Posted : November 30, 2010 4:51 pm
 BigE
(@bige)
Posts: 2694
Registered
Topic starter
 

Interesting... a while ago I used Google Earth to fly me to the lat/lon of my zip. It took me right down town (a few blocks away, maybe a dozen or so) with a big red dot on it. The PO is a few blocks down the street from that. Makes me wonder if there isn't a geo monument down there. I might just have to take a little foot-trip tomorrow or the next day (weather permitting) to investigate.

I remember asking once before on "that other site" about calculating the distance between a set of lat/lons. I was told it involved much calculus and such. NON SENSE. It's real easy - well for someone with a mathematics background/major. Making it into some workable program was not that straight forward.

Fun stuff!!! Especially when it worked on the first shot!!
E.

 
Posted : November 30, 2010 6:40 pm
(@dougie)
Posts: 7889
Registered
 

> Makes me wonder if there isn't a geo monument down there.

More than likely, the Lat/Lon is the center of the area, comprising the zip code;-)

Cheers,
Dugger

 
Posted : November 30, 2010 6:53 pm
(@butch)
Posts: 446
Registered
 

Definitive radius? no such animal

Mean radius = 6,371,000 m per Anderson & Mikhail. Mean radius in conjuction with NAD27 = 6,372,000 m. Your distance in feet (or miles) would depend on if you're using US Survey feet or international foot definition. There shouldn't be more than these 2 approximations for the US anyways.

 
Posted : November 30, 2010 7:31 pm
(@deral-of-lawton)
Posts: 1712
Registered
 

ZIP code centroids are defined by the US Census Bureau and they are a point near the center of the areas of a polygon for a 5-Digit ZIP Code. The centroid is calculated as the internal balance point, based on the coordinate extremes of the polygon. In cases where the polygon is irregular, the centroid may be adjusted so that ZIP Code labels never fall outside of the polygon. So for every single ZIP code, there is a matched ZIP Code centroid value that corresponds to a specific latitude / longitude pair of values.

So, the centroid is not related to a monument, nor a post office but to a mathematical computation based on the extremes of the polygon. These will not often be in the visual center of the zip code area.

I do not use the centroids for distance calcs but use the polygon instead.

Deral

 
Posted : December 1, 2010 2:32 am
(@deral-of-lawton)
Posts: 1712
Registered
 

FAQ on Zipcodes

And this is further information on zip codes in general.

 
Posted : December 1, 2010 2:38 am
 BigE
(@bige)
Posts: 2694
Registered
Topic starter
 

Interes3ting stuff.
I also have a ton of info (and data base) regarding area codes for north America.

 
Posted : December 1, 2010 6:43 am