494 
driven plotters, and new host minicomputers that the 
staff needed to conduct the numerous activities asso 
ciated with building and using the TIGER data base. 
• To build a computer file containing every known 
street and road in the United States (see Notes 1 and 3), 
the name (or names) of each, and the range of address 
numbers located along each segment of every street in 
the 345 largest urban areas of the United States; all the 
railroads in the United States; all significant hydro- 
graphic features and their associated names; named 
landmark areas such as major parks and military 
bases; and essential “key geographic locations” such as 
named apartment buildings, shopping centers, fac 
tories, and office buildings that are important as 
alternate ways to address mail. 
• To enter and verify the boundaries, names, and 
numeric codes for all the geographic entities used by the 
Census Bureau to collect and tabulate the results of both 
the 1980 and 1990 decennial censuses of the United 
States. In doing this, the TIGER data base also includes 
most of the more limited set of geographic entities used 
for the economic and agriculture censuses of the United 
States (U.S. Bureau of the Census, 1985a). (See Figure 1) 
The Process 
The Census Bureau derived the initial set of information 
this data base contains from three primary sources: 
1. The 1980 census GBF/DIME-Files — covering less 
than 2 percent of the land area but 60 percent of the 
people in the United States -- were the primary source 
for the Nation’s major urban areas. The GBF/DIME- 
Files contained the features, feature names, and ad 
dress ranges compiled by the local officials who helped 
the Census Bureau create those files over the course of 
12 years and two decennial censuses — information 
presumed to be basically correct except for changes since 
the last update of the GBF/DIME-Files in 1976. 
2. A cooperative program with the U. S. Geological 
Survey (USGS) -- “the national mapping agency” for the 
United States -- provided the primary source for most of 
the remaining 98-plus percent of the Nation’s territory 
(McKenzie and LaMacchia, 1987). The features in the 
USGS files were compiled to National Map Accuracy 
Standards using aerial photography that was no more 
than 3 years old at the time a particular map was 
prepared — and most were prepared during the 1983-1987 
project period. This was in contrast to the 1980 census 
when the maps used for the 98-plus percent of the Nation 
outside GBF/DIME-File areas were “the best the Census 
Bureau could find” — often state and local maps with 
compilation dates 10, 20, or more years earlier. 
3. For Alaska, Hawaii, Puerto Rico, and the other areas 
for which the Census Bureau created the TIGER data 
base (see Note 1), the Census Bureau digitized available 
maps — typically published USGS quadrangles ranging 
in scale from 1:20,000 for Puerto Rico to 1:250,000 for the 
remote areas of Alaska. These maps varied widely in 
age. 
To make the information in the initial TIGER data base 
more current for 1990 census operations, the geographic 
staff in the Census Bureau’s 12 regional offices collected 
the latest available maps from local officials across the 
United States. The geographic staff then compared the 
information shown on those local maps with the 
information contained in the developing TIGER data 
base. Where there were differences, they inserted new 
streets and roads, and appended street and road names 
to those new features when the local maps showed 
names for them. The geographic staff also used those 
locally collected source maps to append names to all the 
streets and roads that entered the TIGER data base from 
the USGS files; those files came to the Census Bureau 
with no names. This work was done over a period of 4 
years, using quality control checks designed to keep the 
number of clerically induced errors at “normal” levels 
(Marx and Saalfeld, 1987). 
This development task is complete; the goal of 
supporting the 1990 census, nearly met. The resulting 
computer file contains a latitude and longitude 
coordinate value for each of the more than 28 million 
feature intersection and end points that define the nearly 
40 million feature segments that outline the approx 
imately 12 million polygons in this giant "connect-the- 
dots" map of the United States. The foregoing leads, 
more logically at this point, to a discussion about 
geographic information systems (GIS). 
GIS APPLICATIONS 
What is a GIS? The debate goes on endlessly. Some 
argue that the only “true” GIS is a computer system that 
processes “overlays” of information. Others contend that 
a GIS must include natural resources data. Still others 
debate the level of accuracy one needs in the coordinates 
that define the features and data polygon boundaries 
comprising the GIS. Without endorsing or excluding 
any of these definitions, providing a more generic defini 
tion may help the average citizen understand the value 
of a GIS. The more generic definition is: 
A computer system that helps people discover 
relationships between and among sets of geographi 
cally-referenced data that they could not see or 
understand easily without the aid of this technology. 
Whichever definition one accepts, the facts show that 
GIS technology is helping the United States and many 
other countries around the world study the numerous 
environments and changes that surround us. These 
enormous amounts of money and energy are being spent 
under the rubric of “global change.” 
In the more generic context suggested above, it seems 
that the users of most GIS products available thus far 
have focused on physical features and natural resources 
data. In doing so, they typically have ignored one other 
important factor that surrounds us — people! 
People in a GIS 
People inhabit almost every part of the earth. They 
either affect what is going on or are affected by what is 
going on. This is especially relevant if one accepts the 
predictions of many demographic experts who project 
that in the next 50 years, the earth’s population will 
double (Torrey, et al., 1989). They also project that this 
growth will continue the dramatic shift from rural 
settlement to urban settlement with the urban pop 
ulation likely to triple in the same 50 years. Certainly 
these shifts are having — and will continue to have — a 
significant effect on the world in which we live. For this 
reason, it is desirable that people be a “layer” in a GIS. 
Perhaps the omission of “people” information from most 
GIS studies has resulted because this “people” infor 
mation traditionally has not been in a form that com 
puters could process conveniently in a GIS context. 
With natural resources data from satellite imagery, the 
opposite has been true. It streams down from the sky 
with, as Carl Sagan would say, its "billions and billions” 
of pixels. The pixels in a satellite image basically 
portray little polygons or rectangular sections of the 
earth’s surface. The resolution of those pixels varies 
from “coarse” to “fine,” depending on the sensitivity of 
the instruments in the satellite transmitting them.