CDP's do not represent jurisdictions that are incorporated, and are not meant to follow other legally defined entities, so I am not following your take on the correspondence or lack of it between CDP boundaries and other local boundaries. The area may have changed between the time CDPs were delineated and the census date, so that a CDP might not represent the full extent of a community at the time of the census. The one CDP boundary update that is and must be made just before the census is to trim CDPs to align with corporate limits of adjoining legal places as of Jan. 1 of the census year. Also, if you are looking at the cartographic boundary files, they are all generalized to simplify their shapes, and reduce their size and drawing time, so they no longer match exactly to similar files taken from the TIGER/Line Files, which show every vertex or shape point we have stored for the boundaries.
The features that are getting updated presently in the MAF/TIGER accuracy improvement part of the enhancement are primarily roads and streets, with some hydrography and a few local boundaries. To get the best data sources and further partnerships, local/tribal government sources are being used wherever possible. The results of early coordinate improvements for 20 counties were in the 2002 TIGER/Line Files and 107 more counties will show up in the 2003 files.
CDP boundary updates are still on a decennial schedule, right now, although census planners are looking at how they might be revised more frequently to serve the ongoing American Community Survey, since it will provide updated statistics every year. For the areas below 65,000 population, like most CDPs and incorporated places, it will take 3 to 5 years to accumulate enough of a sample for reliable estimates of population and housing characteristics and the statistics will be rolling averages.
Another part of the enhancement program is redesigning the whole census geographic support system, and one goal of that is to make the information more easily shared and updated with our partners. Technology and standards are both headed in the direction of interoperable systems that will make it much easier to use one another's data without import/export transformations, and we are actively working with Open GIS Consortium and others to make that happen.
Charles Dingman, Geographer/301-763-1120 Rm 425, WP-1, Geography Division US Census, 4700 Silver Hill Rd Stop 7400 Washington, DC 20233-7400
|---------+----------------------------> | | "Robert Heitzman"| | | <rheitzman@hotmai| | | l.com> | | | | | | 10/27/2003 11:34 | | | AM | | | | |---------+----------------------------> >--------------------------------------------------------------------------------------------------------------------------| | | | To: charles.e.dingman@census.gov | | cc: CWeaver@icfconsulting.com, gislist@lists.thinkburst.com | | Subject: Re: [gislist] Census Boundary Files Origin | >--------------------------------------------------------------------------------------------------------------------------|
>I am not sure what "populated places" are meant, because I don't see >anything called that in Census products. The data sets Caliper puts >together include census places, legal and statistical, as polygons, from >Census sources, along with point locations of "populated places" from USGS >sources, the GNIS I believe, a data base of geographic names. The boundary >files Census distributes come from the detailed geographic data base used >in the Census operations and tabulations, TIGER, as do the TIGER/Line >Files. Places in that context mean CDPs (statistical areas) and >incorporated places (legal areas).
I stand corrected, I mixed terms from the two sources.
> >Incorporated places and other legal areas come from the Boundary and >Annexation Survey returned by local officials as the boundaries in effect >Jan. 1 of the census year. The legal limits do not have to follow any >visible feature, and are always made census block boundaries. CDPs are >defined before the census in cooperation with local officials, often >planning boards, and have to follow visible features, with a few exceptions >such as short extensions of streets into RR or river features, and >point-to-point lines between physical features. They are defined to >
|