[Talk-us] Facts about the world

TC Haddad tchaddad at gmail.com
Sun Apr 5 02:00:19 UTC 2015


On Sat, Apr 4, 2015 at 2:21 PM, Darrell Fuhriman <darrell at garnix.org> wrote:

TIGER is much older than reasonably portable GPS units.


According to [1] pre-TIGER paper map sources were below the quality of the
1:100,000 DLG data outside of urban areas. Many of the crazy spaghetti data
areas we see probably haven't changed geometry in TIGER much since since
the mid-1980's when they were digitized the first time. They were never
intended for mapping, but rather were "attribute rich" for relative
geocoding. We shouldn't be surprised they as bad as they are. Many of the
processes for rural counties to contribute modern geometry (not attributes)
to TIGER only gained steam after the date that OSM sucked it it's first big
import, so the bad areas never benefitted from the geometry improvements.
It is what it is...

All this is to say is that it's important to understand the origins of any
data - many of the complaints people have about government data sources are
easily explained if you trace the origins. For example many people don't
know that the original source of the digital spatial coordinates of
uncorrected GNIS point data was the centroid of the cartographic labels on
24k topo quads. This entirely explains why the points are most often next
to, rather on top of the feature being labeled. Yet complaints abound
because for OSM scales, the data is imperfect. So it goes.

[1] http://geospatial-solutions.com/tiger-database-historical-perspective/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20150404/220fd9b7/attachment.html>


More information about the Talk-us mailing list