[Talk-us] musings on TIGER tags

Nathan Edgars II neroute2 at gmail.com
Tue Nov 23 10:24:54 GMT 2010

tiger:cfcc: "census feature class codes used to differentiate feature
types"; describes the importance (generally primary=U.S. Highway,
secondary=state highway) and type of road
Accuracy: low. Alignments of main routes through towns almost never
take the proper turns. Bridges and tunnels are often way too long.
Usefulness: nil. We have other tags for this that get updated and corrected.

tiger:county: the county in which the feature lies
Accuracy: high, though may have issues right at county lines.
Usefulness: not sure. How much is is_in used on roads worldwide? Could
be calculated from county polygons (if TIGER's polygons are imported
for all states).

tiger:name_base, tiger:name_direction_prefix,
tiger:name_direction_suffix, tiger:name_type, tiger:name_base_1, etc.:
different parts of the street name
Accuracy: relatively high, with problems primarily when numbered
routes are included.
Usefulness: not directly as a TIGER tag that doesn't get updated and
corrected when the name field is. Could be useful as a non-TIGER tag.

tiger:reviewed: default no, changed or removed when someone has reviewed the way
Accuracy: perfect at the time of import.
Usefulness: very useful.

tiger:separated: indicates whether a dual carriageway is shown as a
single or double line
Accuracy: initially perfect with respect to whether the data used a
single or double line, but not great with respect to whether a road is
dual carriageway in the first place.
Usefulness: nil. A general tag saying whether a one-way road is half
of a dual carriageway might be useful, but this doesn't help.

tiger:tlid: "TIGER/Line Identification Number", a unique (?) number
for each segment inventoried by TIGER
Accuracy: perfect at time of import, but has gotten horrible with
joining and splitting ways. The field length can easily exceed the
API's length limit, leading to loss of information (Potlatch just
truncates, I believe, while JOSM throws an error and leaves the user
to deal with it).
Usefulness: might be useful as a cross-reference to other datasets,
but the low accuracy may make this impossible.

tiger:zip_left, tiger:zip_right, tiger:zip_left_1, etc.: presumably
the ZIP Code Tabulation Area in which the road lies.
Accuracy: initially perfect with respect to ZCTAs, but not actual ZIP
Codes, since ZIP Codes don't work this way:
http://www.census.gov/geo/www/tiger/tigermap.html#ZIP In addition,
some ZIP Codes are not represented by ZCTAs at all, such as 32830
(assigned to Walt Disney World since 1971).
Usefulness: low due to the problems mentioned above, as well as
degradation caused by joining and splitting ways. Could be calculated
from ZCTA polygons by the end-user.

tiger:source, tiger:upload_uuid: appears to have been added by the upload bot
Usefulness: low to nil. A source=TIGER shapefiles etc. would seem to
be just as useful.

More information about the Talk-us mailing list