Information in this description is drawn from the TIGER/Line(TM) file technical documentation[1]. The full document, in Adobe PDF format, along with sample data sets and other useful information, can be found at the TIGER home page. Here is a local copy of the Redistricting Census 2000 Tiger/Line File Technical Documentation. (This is the version of the data that we are using in our Map Server Demo.)
TIGER stands for Topologically Integrated Geographic Encoding and Referencing. The TIGER system was developed by the U.S. Bureau of the Census. The Bureau provides the public with the TIGER/Line files, which are ``extracts of selected geographic and cartographic information'' from the TIGER data base. Among other things, the Bureau also provides TIGER/SDTS(TM), which is ``a relational data file following the FIPS [Federal Information processing Standards] SDTS [Spatial Data Transfer Standard]''. This file contains ``data equivalent to the TIGER/Line files with additional relational data linkages and data content more similar to the Census TIGER data base.''
The TIGER/Line files are county-based, and they can be combined to cover the entire United States and its territories. They contain three types of data: line features, such as roads and railroads; landmarks, such as schools and parks; and polygon information for area boundaries. The line features and polygon information form the bulk of the data.
All latitude and longitude coordinates in the Tiger/Line files are
given to six implied decimal places. The field width for latitude is
9 characters, including the sign (+/-), and the field width for
longitude is 10 characters. Hence, the longitude values have room for
three digits to the left of the (implied) decimal point.
For example, ``-101795270+44090946'' represents the point
with longitude -101.795270 and latitude 44.090946. The sources
for the TIGER data include the USGS 1:100,000-scale Digital Line
Graph, so the ``positional accuracy of the information is no greater
than the established National Map Accuracy standards for
1:100,000-scale maps from the USGS (approximately +/- 167 feet)''.
However, relative positions of the elements should be correct.
The TIGER data contains three kinds of geometric objects: points, polygonal lines, and polygons. The polygonal lines are called "complete chains". Each point is represented by a latitude/longitude pai, each complete chain is represented by a sequence of points, and each polygon is bounded by a sequence of complete chains. In addition, there is voluminous qualitative information associated with the geographic features, such as zip code, address range, etc. The following section gives a brief summary of how the data is stored in files.
The data for a county may use up to 17 different record types, and all the records of a particular type are grouped together in a file. Not all record types will necessarily be used for every county. The Tiger/Line files are all ASCII, and each line represents one record, so the end-of-line character is also the end-of-record character. There are a constant number of fields for each record type, and each field in a record is a constant number of characters. If a field is missing, it will be filled with blanks (or zeroes, in the case of record type 2). What follows is a brief description of each record type.
Record types 1 and 2 can be used to piece together all of the polygons in the county, and much of the other information as well. Each polygon is composed of one or more ``complete chains'', where each complete chain is a polygonal line. Complete chains may intersect only at their endpoints. Each complete chain serves on the border of two neighboring polygons, except for rare ``dead end'' chains, which have the same polygon on both sides. The endpoints of a complete chain are specified in record type 1, along with a unique Tiger/Line ID, or TLID. If the complete chain is not a straight line, then its intermediate points are specified in one or more records of type 2, which are linked to the chain's (single) record type 1 by the TLID. The only purpose of record type 2 is to store intermediate points, so not all chains need type 2 records.
In addition to the polyline endpoints, record type 1 contains many descriptive fields, including,
Road, separated, with rail line in center, and H22 means
Intermittent canal, ditch, or aqueduct);
Here is a sample record of type 1, except that the spaces have been
replaced with .
characters, and the single line has been split
into three lines, each ending with a \
character, to fit onto
the page.
10024..75598596.BN.Front.........................St....A31.........40.......\ ..42.........45.........4700010435704357............4023230230236264562645..\ ..................9801..9801..243.250..-69795270+44090946.-69794636+44091152\
The first five characters give the record type (1) and version
number (0024). The next ten characters represent the
Tiger/Line ID, which is unique for each record of type 1. The
following character holds a flag which indicates whether the
line is single-sided. A single-sided line lies on the county
boundary, so that the polygon on either its right or its left side is
outside the county. In that case, certain fields, such as census
tract and census block number, would be left blank for the side that
is outside the county. The blank indicates that this line is not
single-sided. The next character,
B, is a code for the source of the data. Code B
indicates USGS 1:100,000-Scale DLG-3 File
. This is followed by a
feature direction prefix (2 characters), a feature name (30
characters), a feature type (4 characters), and a feature direction
suffix (2 characters); i.e., N. Front St. The three character code
A31 is the census feature class code for Secondary and
connecting road, State and county highways, unseparated
. This is
followed by four fields of eleven characters each representing the
start and end addresses on either side of the line. The next four
characters are impute flags
for each address range endpoint. The
impute flag indicates that an address range is based on calculations
rather than known values. This can happen when a Tiger/Line is split
(by a new road, for example), and it is not known at which
point the address range was split. The next ten characters give the
ZIP code on either side of the line (04357), and the ten characters after that
(here blank) give codes for American Indian/Alaskan Native areas. The
next two characters are flags for American Indian Trust Land, left and
right side. And the two after that are for census internal use.
This is followed by FIPS state code on either side (two characters
each; 23 represents Maine), FIPS county code on either side (three
characters each), and
other FIPS codes for subdivisions and places (thirty characters
total). Next are six characters each for the Census Tract number on
either side (9801..), and four characters each for the Census
Block number on either side (243. and 250.).
Finally,
the longitude and latitude coordinates for the start point and end
point of the line. These coordinates have an implied decimal point,
with six digits to the right of the decimal point. The longitudes
have space for three digits to the left of the decimal point; the
latitudes only need two.
Here is a sample record type 2, formatted in the same way.
20024..75605696..1.-69890008+44126686.-69890516+44126343.-69890897+44126183\ .-69891595+44126229+000000000+00000000+000000000+00000000+000000000+00000000\ +000000000+00000000+000000000+00000000+000000000+00000000\
Record type 2 also starts with the record type, version number, and
Tiger/Line ID. This is followed by a three character field giving a
sequence number
, and ten fields for longitude and latitude of
shape points. If fewer than ten shape points are needed, the
remaining fields are padded with zeroes. If more than ten shape
points are needed, then multiple records of type 2 will be given for
that Tiger/Line, and numbered with the subsequent sequence number. The
record shown has four shape points, and it is sequence number 1.
Since there are no more shape points for this line, it will be the
only record of type 2 for the line. A full line is a polygonal line from the
start point in the record of type 1, to the shape points from any
records of type 2, in order by sequence number and in order within the
records, and finally to the end point from the type 1 record.
As an example, Sagadohoc County, Maine, contains 7540 type 1 records, and 5875 type 2 records.
The Census Feature Class Codes, which are used to describe line
features in record type 1 and landmark features in record type 7,
consist of three characters: a letter which identifies the feature
class, a number which describes the major category, and a number
describing the minor category. For example, the feature class code
for the example line above is A31. The feature classes
are
Features that are unknown or not yet classified are given a code beginning with X. Either of the major or minor categories may be given a digit 0, indicating that the category is not known. This document describes the road feature class in detail, and the other classes cursorily.
There are seven major categories associated with the road feature class:
Secondary and connecting road, State and county highways, unseparated, and A18 represents
Primary road with limited access or interstate highway, separated, with rail line in center. The
rail line in centerminor categories include any situation where the railroad and road are both represented by one complete chain, even if the railroad is next to the road and not
in center. None of the
separatedminor categories exist for the vehicular trail major category.
The last two major categories each have special minor categories.
The road with special
characteristics
category includes cul-de-sacs, traffic circles,
access ramps, service drives, and ferry crossings. The road as
other thoroughfare
category includes walkways, stairways, and
alleys.
The railroad feature class includes the following major categories: railroad main line, railroad spur, railroad yard, railroad with special characteristics, and railroad as other thoroughfare. It includes the following minor categories: not in tunnel, in tunnel, and underpassing.
The miscellaneous ground transportation class includes pipelines, power transmission lines, tramways, monorails, and ski lifts.
The landmark feature class includes military
installations, multihousehold or transient quarters, custodial
facilities, educational or religious institutions, transportation
terminals, employment centers, towers, and open space.
Feature class codes for landmarks include, for example, D81 for Golf
course
and D36 for Jail or detention center
.
Physical features include fences, ridges and mountain peaks.
Nonvisible features include such things as legal or administrative boundaries, property lines, and ZIP code boundaries.
The hydrography feature class contains the following major categories: naturally flowing water features (rivers and streams), man-made channel to transport water, inland body of water, man-made body of water, seaward body of water, body of water in a man-made excavation, nonvisible definition between water bodies, and special water features (e.g., glacier).
1. US Department of Commerce, Economics and Statistics Administration, Bureau of the Census, Washington, D.C. TIGER/Line(TM) Files, 1995 Technical Documentation, 1996.