Bacon's Rebellion

Democracy Thrives in Sunlight

ZIP CODES AS DATA SOURCES

—

by

James A. Bacon
in

Months ago EMR promised Larry Gross a response on the issue of Zip Codes as a source of information. Here is an excerpt from Chapter 35 in PART TEN of TRILO-G:

…………………

Data from Postal Zip Codes provides an excellent example of what is wrong with current data resources. This post might be labeled “More on the Problems of Zip Cones.” Readers can find 21 references in The Shape of the Future to the problems caused by Zip Cones when one is trying to understand human settlement pattern.

Zip Code data is “interesting” BUT…

Recently a person who comments often on the Bacons Rebellion Blog emailed the author information on a web site that he assumed would help answer settlement pattern questions: ZipWho.com.

The site is a useful one. However, just one example – that was typical of several trial comparisons by the author – documents the pitfalls of Zip Code data.

…….. the example is included as the first comment on this post ………

If one understands human settlement patterns and has other sources of information as well as accurate maps, the ZipWho.com data can be useful. Without this understanding and additional information, the data is at best deceptive.

As noted in The Shape of the Future, Zip Codes are NOT organic components of human settlement. They vary widely in size and the configurations reflect the distribution of power among appointed Postmasters when Zip Codes were established and when they are revised.

EMR

Share this article

ADVERTISEMENT

(comments below)

ADVERTISEMENT

(comments below)

Comments

14 responses to “ZIP CODES AS DATA SOURCES”

  1. E M Risse Avatar
    E M Risse

    Here is the example from the text of Chapter 35:

    ZipWho.com provides 17 categories of information on every Zip Code in the US of A. Further it provides the national percentile rank of each of the 17 categories and highlights those categories with percentage rankings that fall in the top ten percent or in the bottom ten percent nationally for all Zip Codes.

    For example, on first blush a Median Income ranking in the top ten percent and a Divorced Persons ranking in the bottom ten percent might suggest a “good place to live.”

    The author considered two Zip Codes 20186 and 20187.

    Zip Code 20186 covers the Town of Warrenton and territory to the north, west and south. The area includes the Urbanside focused on the Town, the logical location for the Clear Edge to the north, west and south of the Town as well as “a good approximation of” the Countryside half-way to the next post office to the north, west and south. The population of the Zip Code is 10,588, about twice the Town population. The Town of Warrenton is the County Seat of Fauquier County and is a picturesque “Main Street Community” with an array of commercial establishments to provide services for the residents of the Urbanside and the surrounding Countryside. There are a number of jobs in 20186, although not enough to support the working population or to achieve Balance.

    The 17 indicators for Zip Code 20186 are about what one would expect. If there is to be Balance few would be in the top ten percent or the bottom ten percent of most categories.

    The border of Zip Code 20187 follows the eastern border of the Town of Warrenton and extends north, east and south of the Town with somewhat more territory than 20186. The logical location for the Clear Edge around eastern part of the Core of Greater Warrenton falls in 20187. US Route 29 runs northeast / southwest through 20187 and is the direct route to the I-66 / VRE corridor and to the Core and Centroid of the National Capital Subregion.

    Most of the 20187 territory has been a Census designated “Urbanized Area” for two decades due to the scattered orphan subdivisions. The Urban Dwellings in this Zip Code are regularly advertised with the Real Estate Ad moniker “on the DC side of Warrenton.” There is a small amount of strip commercial along US Route 29 but no concentrated area to acquire services. There are a few jobs but nothing like the number needed to support the 13,288 residents.

    The 17 indicators for 20187 are not surprising but the could be misleading. There are six top ten rankings (Median Income, Owner Occupied Homes, Median Rooms in Home, Percentage of Professionals and Average Household Size) and one bottom ten ranking (Divorced).

    The settlement pattern for 20186 is far more functional but that could not be determined from the data. One needs to look at the cost of transport and housing and find that along with a lack of Balance there are higher costs based on Census Block Group data. See Blog Post “Back to ‘Beltway Burden,’” 15 February 2009 in PART FOURTEEN – RESOURCES.

    If one understands human settlement patterns and has other sources of information and accurate maps the ZipWho.com data is useful. Without this understanding and additional information, the data is at best deceptive.

    As noted in The Shape of the Future, Zip Codes are NOT organic components of human settlement. They vary widely in size and the configurations and reflect the distribution of power among appointed Postmasters when Zip Codes were established and when they are revised.

    EMR

  2. Larry G Avatar
    Larry G

    perhaps what is known as “carrier routes”?

    “Is there a unit smaller than the ZIP Code?
    Yes. To aid in mail delivery, the USPS divides ZIP Codes into carrier routes. As the phrase implies, these are determined as the actual physical area an individual mail person can cover every day. Carrier routes always sync within their corresponding ZIP Code.”

    http://www.maponics.com/ZIP_Code_Maps/ZIP_Code_FAQ/zip_code_faq.html

    Carrier Route Maps and Carrier Route Mapping Services
    Maponics has been working with the US Postal Service to provide ZIP Code and carrier route maps to businesses since 2003. Each year, thousands of direct marketers rely on us to help them accurately target prospects by mail.

    http://www.maponics.com/Carrier_Route_Maps/carrier_route_maps.html

    I think… EMR.. there is probably sufficient raw data… to use for your purposes but you’d probably have to hook up with someone who is good at slicing and dicing data….

    For me… I’d not need to see a complete and comprehensive map of finished data…

    but perhaps one example… say pick a location in NoVa … and show us your 10x rule.. and then overlay it with the zip (carrier?) codes to show how your template would differ.

  3. Anonymous Avatar
    Anonymous

    EMR is right. Zip codes are a lousy way to collect data. They might not even be an efficient way to deliver mail.

    When was the last time your zip code changed?

    We probably should create an algorithm that assigns legislative districts by population in the most efficient and compact way. This would change over time with population shifts, but in a uniform way.

    Such a distribution should meet most of EMRs criteria,,and we could collect data by legislative district.

    We could have more legislators representing smaller and more localized groups, too.

    RH

  4. Larry G Avatar
    Larry G

    I’m not sure that doing the data by existing legislative boundaries will suit EMR as he often talks about the need for a change in governance.

    To be honest, I’m not sure what EMR’s criteria would be but …as I pointed out.. each Zip code consists of carrier routes – which currently are much smaller in size than the Zips and … look like they are according to streets….”carrier routes”.

    I’d say the demographic data … exists at pretty granular levels .. if the marketing folks are using the carrier routes for direct mail – as the provided link confirms.

    I think EMR could make some headway …. if he could demonstrate his boundaries by using them as a template for zip codes or carrier groups and this would give the average person a much, much better idea … a graphical picture (with at least a 1000 words!).

  5. Anonymous Avatar
    Anonymous

    It would not be the existing legislative boundaries.

    They would be new ones created by some new politically independent algorithm. It would be defined throughthe political process however so EMR would have some kind of input.

    He still won’t be satisfied withthe results, but…..

    I’d suggest a mixed metaphor algorithm that starts making new boundaries starting with the smallest and densest districts and balancing that against some other apolitical criteria like starting at the population center of the sate and working out, or northwest corner of the state and working towards the south east.

    The idea might be something like “Starting here, what is the smallest six sided polygon that will include 33,000 people,the number required for a legislative district.

    No more gerrymandering, and no more engineering safe districts. By searching for the smallest possible districts in any case and the very smallest ones first, you might overcome some of EMR’s objections.

    RH

  6. E M Risse Avatar
    E M Risse

    EMR will get to Larry soon but first:

    EMR has requested release from a funeral-day promise to Hon. Lynn hopewell…

    This is the first response to RH that EMR has made since that promise was made.

    EMR is not responding because RH suggested that EMR was right.

    He is responding because RH is right on all counts.

    Zip Code data has been crafted for mail solicitations. That may work well but other than that…

    RH does not know it but his “mixed metaphor algorithm” could be an element of a larger system to create a new set of borders that EMR has already sketched out.

    RH would be surprised at how many elected officials are aware of the pitfalls of current borders but they do not know how to get off the Tiger.

    When EMR has completed TRILO-G he has a commission to do a pilot project on this topic. If any of RH’s ideas are used he will get royalites. EMR beleives that will suit RH’s view of intelectual property rights.

    I think Lynn would approve. Fundamental Transformation of governace structure is the sort of thing he would have been interested in.

    EMR

  7. E M Risse Avatar
    E M Risse

    Larry said:

    “I think… EMR.. there is probably sufficient raw data…”

    Right you are! There are mountains of RAW data.

    The firm to which you provided links above will be happy to sell you data. The data is already aggregated for mass mailing purposes so it will cost you to disaggregate it and that may or may not be possible.

    Selling data is also what the creator of ZipWho.com does – in addition to collecting live e-mail addresses for marketing.

    They are paid to answer the question: “Where do the folks live that might buy product X or service Y?”

    That has little to do with creating Balance as I tried to point out in the post.

    A place with Balance would have a spectrum of incomes, lifestyles and activities. Nature abhors monocultures (and vacuums) but Marketing LOVES monocultures.

    The focus on “who we can sell to” is a reflection of the culture that drives Mass OverConsumption.

    The Census Bureau has mountains of RAW data too.

    When Lucy and Phillips first came to EMR with the preliminary findings that become their important book “Confronting ‘Sub’Urban Design” the conclusions were based on Census Data.

    EMR pointed out that their conclusions did not match reality because the data aggregations were not organic components and thus mixed a number of generations and types of development which they were trying to focus on.

    EMR also told them that they were onto something and suggested they go back and get data from smaller Census defined geographies and then reaggregate the data by smaller components to get something approximating Neighborhoods. They did that – as described in their methodology – and the rest is history.

    They used the same techniques in their later book “Metromorphosis” which later had its title changed for marketing reasons. These efforts took them thousands of hours, lots of grad students and hefty computing capacity.

    EMR has never said there are no sources of data or that in its RAW state the data is not valuable. What EMR has said is that when aggregated by arbitrary geographies – municipal borders, Zip Codes, etc. – it is of little use.

    “to use for your purposes …”

    EMR is not sure what you think those ‘purposes’ might be.”

    “but you’d probably have to hook up with someone who is good at slicing and dicing data….”

    Any large, well funded sturdy could acquire those services, especially now.

    “For me… I’d not need to see a complete and comprehensive map of finished data…”

    Then your budget need not be overwhelmingly large.

    “but perhaps one example… say pick a location in NoVa … and show us your 10x rule..”

    You lost EMR there. EMR already provided you with all the framework information you need to do a proof of the 10 X Rule.

    The data you need is information on actual costs of delivering 40 + / – location variable goods and services.

    Since the 10 X Rule is based on a hypothetical 10,000 acre area with alternative settlement pattern distributions using the same per foot and per hour unit costs, EMR has no idea how Postal or Census Data would help in your effort but that is for you and your team members to work out.

    Good luck on your project!

    There is a larger issue here. By mixing up the data that can be gleaned from Zip Codes, Census, IRS, BLS, etc. with the data needed to prove the 10X Rule or create Balance Larry demonstrates both the complexity of the task and the weakness of a Blog to communicate about those tasks.

    EMR

  8. Larry G Avatar
    Larry G

    ” governance per the 10x rule”

    how are wards.. district boundaries determined?

    In my county… density ends up having an effect because of the one-man, one-vote rule.

    so the denser an area is – the smaller the representative boundaries at the lowest governance level – the Supervisor.

    I’m not sure how these districts get changed … but they do so .. about every ten years when the new census data comes out.

    by the way.. that’s also the same event used to determine MSAs.

    Zip Codes can.. and are changed as well as the location name.

    One of the problems in the Richmond to NoVa corridor has been changing he zip code but keeping the location name.

    For instance, the core Zip Code for Fredericksburg is 22401.

    As the area has grown.. new zips have been added and now we have a dozen zip codes all with Fredericksburg as the mailing address.

    Problem is.. these are Spotsylvania addresses… and… believe it or not.. the state’s allocation of things like sales taxes and other taxes (and services) has been found in error.

    Henrico County has had the same problem…. they found their sales tax revenues being given to Richmond just as Spotsylvania found theirs being given to Fredericksburg.

    Find and fixing this has caused some major wailing and gnashing of teeth… as the net result for Richmond and Fredericksburg is a net loss of tax money.

    this fact – that localities depend on the revenues from sales taxes – has wreaked havoc on land-use and transportation planning in Va – and in turn.. have been not helpful with settlement patterns….

    because.. localities “compete” for commercial tax revenues.. there are permanent “wars” over who can attract commercial businesses and this is how we end up with a lot of “left over”.. semi-abandoned huge-parking lots that end up with a bunch of low level commercial like tanning parlors and baseball card sales, etc.

    Way..way too much “commercial” that ..cannot be easily used for “new urban” without a willing investor and a strong willingness on the part of the locality to remove the roadblocks from more dense residential re-development.

    I’ve always found that a tad bit curious… BR has laid down multiple threads on “mixed use” development.. but invariably they are “greenfield” development… and sometimes surrounded by defunct shopping venues…

    One of the ones talked about in BR… is on the “outskirts” and you have to literally drive past about a 1/2 dozen older.. semi-decrepit retail venues to get to the “new” planned development.

    Finally.. I’m still a major skeptic that we can …”create” or even modify/redevelop existing development … in a world economy where how people are actually able to make a living… is undergoing some serious changes…

    and we all know.. you can have the perfect grid-street “balanced community” that morphs into something no longer vibrant if the jobs go away to somewhere else.

    Settlement Patterns.. dysfunctional or optimized wither away if the jobs go away and don’t come back.

    this IS something that zip codes can and do show.

    It’s called mean and standard deviation – a histogram.

    A histogram not only shows the median income of an area – it shows the top and bottom incomes also and it shows the distribution.

    So a place like Great Falls is going to be a narrow spike.. essentially… and a place like Detroit… especially a “before” and “after” chart is going to show a dramatic difference … a “widening” of the distribution and a “flattening” out of the highs and much more narrow peaks.

    I’d like to see a histogram of a well-balanced community EMR.

  9. Anonymous Avatar
    Anonymous

    What have I got to do with Lynn Hopewell? As far as I know we have never met. Don’t think I ever heard of him before.

    RH

  10. Anonymous Avatar
    Anonymous

    The focus on “who we can sell to” is how we make our living.

    How is your book sale doing?

    RH

  11. Anonymous Avatar
    Anonymous

    I observed some time ago that RH had probably insulted so many people at public meetings that he had no idea who they were.

    Turns out I was right.

    Dr. Risse was sure, given Hopewells intense dislike for RH that they had a real knock down drag out.

    Turns out Professor Risse was wrong.

  12. Anonymous Avatar
    Anonymous

    “I observed some time ago that RH had probably insulted so many people at public meetings that he had no idea who they were.”

    If you insulted someone at a public meeting, how WOULD you know who they were?

    By my recollection, I may have actualy spoken at between four and six public meetings. If that observation is correct it must be some kind of record to aggravate so many people.

    On the other hand, people who are emotionally invested in some kind of belief tend to be highly and irrrationally upset when you present them with an obvious and incontrovertible truth.

    For example, I went to a meeting that involved a couple that wanted to turn an old mansion into a bed and breakfast. A swarm of people went to the microphone one after another to state that Rokeby Rd was too narrow, dangerous, curvy, rough, and populated with horses to allow any more traffic.

    After a steady diet of this, I eventually took the microphone and observed that I didn’t know anything about the bed and breakfast, nor had any feeling one way or another. But that, based on what I had heard it was pretty obvious that Rokeby road needed to be improved, regardless.

    That pretty much brought the house down, but it must have insulted some people because the next thirty minutes was filled with the same stream of people returning to thes mike to explain why Rokeby Rd SHOULD NOT be fixed.

    I thought it was an object lesson in mass hypocrisy, but nobody else seemed to “get it”. The proposal got turned down,although it was a good one and the couple had worked hard to get community buy in. Then they went to the meeting and got ambushed.

    I didn’t expect to change the outcome: that was a foregone conclusion. All I wanted was for people to look themselves in the mirror.

    I guess they were incensed with me because they didn’t like what they saw. Especially since they had no reason to attack me: I no stake in the fight.

    ————————–

    More likely, people know of me from a long string of letters published in the Citizen before it was bought out by the Democrat.

    I’ve only had one letter published since, some people beklieve it is because of the purchase by a group with different politcal views. I’m willing to let them think that, but the truth is I’ve only sent in one letter.

    It is too bad that some people take such umbrage at the exercise of free speech.

    Shucks, it is only a letter, fire one back. People misunderstand me: I like green and clean as much or more than anybody. It was my business before I learned how much I disliked green dishonesty.

    I’d go back in a heartbeat if I didn’t think that, as presently practiced, it involves stealing. I’d go back if I thought I could do something to fix it, but shucks, I can’t even convince Larry what the problem is.

    Meanwhile, I prefer to work on something I can succeed at: even farming is easy by comparison to promoting environmental equity.

    RH

  13. zwr Avatar
    zwr

    You can have a look at that site http://www.datanumen.com/azr/ ,where you can find somtway to fix your corrupt or damaged zipfiles.

  14. Anonymous Avatar
    Anonymous

    Kahn, Matthew E. and Vaughn, Ryan K. (2009) "Green Market
    Geography: The Spatial Clustering of Hybrid Vehicles and
    LEED Registered Buildings," The B.E. Journal of Economic
    Analysis & Policy: Vol. 9 : Iss. 2 (Contributions), Article
    2. DOI: 10.2202/1935-1682.2030 Available at:
    http://www.bepress.com/bejeap/vol9/iss2/art2

    This paper uses zip code level data to investigate the spatial distribution of two major "green" products. Using data from California, we document where hybrid vehicles cluster within the state. Using data for the entire nation, we study where LEED registered buildings cluster. By creating a novel measure of community environmentalism based on revealed preference political data, we document that green products cluster in environmentalist communities.

    Well, Duh.

Leave a Reply

Read More