On The Nets: Making Sense Of The Census

by Greg R. Notess

DATABASE, February 1995
Copyright © Online Inc.

----------------------------------------------------------------

     One of the most common statistical questions, across all types of 
libraries, is the population question. How many people are in the 
nearest town, city, county, or state? Fortunately, the United States 
Census Bureau gathers and disseminates an amazing array of 
statistical reports, estimates, and predictions of the U.S. population. 
And now the Census Bureau is making some of their data available on 
the net. 
     As gopher servers proliferated in cyberspace, localities began 
putting local census statistics on their servers. Some sites provided 
quite detailed population statistics for their state, often pulling 
tables from CD-ROMs with reports from the 1990 Census of Population 
and Housing. This lead to good coverage of recent population statistics 
for some states, while other sections of the country had no substantial 
Internet source for population statistics. The advent of the Census 
Bureau on the Internet has changed that, providing full coverage of the 
U.S. with a variety of statistical reports. The Bureau now offers a 
range of Internet-accessible data sources, including a gopher server, a 
World-Wide Web home page, telnet access to the Census Bureau BBS, 
and an anonymous FTP site. 

POPCLOCK 
     For an excellent example of the kind of instantaneous information 
that the technologies of the Internet offer, try the Bureau's Population 
Clock. The PopClock, as it is known, provides up-to-the-minute 
estimates of the resident population of the United States. This is the 
quintessential online information resource. The PopClock is available 
through a Web browser at http://www.census.gov/cgi-bin/popclock. It 
can also be reached by gopher at the URL of gopher://gopher.census. 
gov: 70/0exec0::/Bureau/Population/ gpopclock or through the 
following menus at gopher.census.gov: 

     3. Enter the Main Data Bank/ 
          13. Population/ 

     To its credit, the Census Bureau provides more than just the 
statistic. Directly below the number is a detailed explanation of who 
is included in the statistic and how the projection is determined. 
While the Bureau has made a great step forward into electronic 
publishing with its Internet presence, it still has a long way to go 
before it catches up to its print publications. By no means are even the 
majority of the census reports and other statistics available on the 
net. 
     For those who use the Internet frequently for ready reference 
requests, knowing what is and what is not available for the Bureau on 
the Net can save a lot of time and frustration. Of the multiple access 
points to the Bureau, the gopher and Web interface are the most user-
friendly. However, the anonymous FTP space and the census BBS can 
provide additional options for access. 

FTP SITE 
     The Census Bureau's anonymous FTP site at ftp.census.gov contains 
many of the files that are also available on its gopher server or 
through its Web home page. In most cases, it is easier to find a 
specific report or statistic through gopher or WWW. However, 
knowledge of the FTP site can be useful as a backup if the gopher 
server is down or if a citation only refers to the FTP location. 
Once connected to the anonymous FTP server, change to the "pub" 
directory. The data files are divided into more than 40 directories. The 
first few listings include the following: 

     Census-BBS/ 
     Govt-Stats/ 
     International/ 
     Press-Release/ 
     Sipp-On-Call/ 
     Stat-Briefs/ 
     We-The-People/ 
     agr/ 
     agriculture/ 
     ces/ 
     construction/ 
     crime/
     .
     .
     .

     Be warned that some directories and subdirectories are empty 
while others contain megabytes of data. When a readme file is 
available in a given directory, read that first to determine the kind of 
data in the directory and the naming scheme. The actual filenames can 
be rather cryptic (care to guess what's in hwts93.dat or c20-
9403.txt?) unless the readme file explains the scheme. 

CENSUS BBS
     The Census Bureau's electronic BBS can be reached by a telnet 
connection to cenbbs.census.gov. As with most BBSs, the first visit to 
this BBS entails a registration process. After answering all the 
questions, the BBS menu will finally appear (Figure 1). 
One of the advantages of the BBS is the message database. For those 
looking to communicate directly with people at the Census Bureau and 
those who want to stay up-to-date with Bureau announcements, the 
Census BBS can be useful. But due to the time involved with the 
registration process and the general sluggishness of running a BBS 
over the Internet, the Census BBS takes the back seat when compared 
with the ease of access to the data via gopher or a Web browser. 

CENSUS GOPHER 
     Figure 2 shows the Census Bureau gopher server (gopher.census.gov) 
using a WSGopher client. An examination of the top level menu (shown 
in the upper left-hand corner) shows that most of the actual census 
data will be found in the Enter the Main Data Bank option. The other 
options either lead to information about the Census Bureau or to other 
Internet resources. 
     The Main Data Bank section (lower left-hand corner of Figure 2) 
provides access to the majority of the available statistical 
publications. However, the names of some of the categories can be 
deceiving. Do not expect to find the full text of comparable print 
resources, at least at this point. Reference chestnuts like the 
_Statistical Abstract of the United States_ or the _County and City 
Data Book_ appear as gopher menu headings, but only a few tables from 
each are included on the gopher. 
     For example, the menu heading Statistical Abstract of the United 
States does not provide access to all the tables in the print or CD-ROM 
_Statistical Abstract of the United States_. It only includes a series 
of tables for state rankings. While these tables do include commonly 
requested demographic information for each state, it is a far cry from 
the wealth of information in the full _Statistical Abstract_. 
     Likewise, the County and City Data Book menu entry only includes 
statistics for cities over 200,000. At least the menu is a bit more 
descriptive in this case, qualifying the menu entry with Top 77 City 
Rankings. Unfortunately, no county level information is provided. 
     Another seemingly promising but deceiving menu option is County 
Business Patterns. Detailed statistics on the number of 
establishments and payroll by SIC are available under this heading, but 
again, only at the national and state levels, not the county level. The 
data are available in both ASCII and comma-delimited formats, but all 
as one big report. There are no entry points for particular states or 
specific SIC numbers, as is common in other reports that the Census 
Bureau makes available electronically. The whole report is provided 
but no detailed access is available. 
     Like the PopClock, press releases benefit from the online advantage 
of the Internet. They can be found under the News Releases Hot Off the 
Press option. These press releases can be an excellent source for 
recent statistics. In addition, they can provide contact information 
that may expand on newspaper reports of the same data. 
     There are many useful statistics under the Main Data Bank menu 
headings. But for those of us who depend on the print publications, the 
limitations in coverage keep the gopher server from being a one-stop 
statistical source. 
     On the other hand, one gopher offering demonstrates the future 
potential of electronic publishing on the net. The menu for We the 
People publications (under Main Data Bank/Documents and 
Publications) presents PostScript versions of this print publication. 
With a client capable of displaying PostScript files (either Xgopher or 
WSgopher with Ghostview), the document can be displayed in fully-
formatted text with illustrations and tables directly on the monitor. 
     There are other categories worth investigating under the Main Data 
Bank heading, but one other useful search category is actually found 
under the About This Gopher Server top-level menu: the Keyword 
Search on Our Entire Server option. Keyword search capabilities on a 
local server (sometimes using the jughead software) are a very useful 
feature of any gopher. On the Census gopher, this option can be used to 
track down specific kinds of statistics. Unfortunately, the 
implementation of the keyword search feature is not always effective. 
For example, a keyword search on a state name does not retrieve all 
(or even most) of the documents or menus with statistics for that 
state. A comprehensive and dependable keyword search of the whole 
Census gopher would be a worthwhile addition to the resource. In the 
meantime, the Keyword Search on Our Entire Server is worth at least a 
quick check. 

WEB SERVER 
     At first glance, the World-Wide Web home page for the Census 
(http:// www.census.gov/) appears to be only a slightly more 
sophisticated front end to the gopher server (Figure 3). There is a 
hypertext link to Main Data Bank, which looks quite similar to the 
gopher Main Data Bank. News Hot Off the Press is available on the top 
of the page, followed by the familiar PopClock. Not too far below 
comes the heading Learn more about us, which has much of the same 
information about the Census Bureau as is found on the gopher server. 
     A closer look at both the home page and the Main Data Bank link 
shows that there are significant differences in content between the 
WWW page and the gopher server. True to its multimedia capabilities, 
the Web page offers audio and visual information in the form of 
recordings of radio broadcasts and a trip through the Census Bureau art 
gallery. (No, it is not a collection of Renoir or Monet, but some gifs of 
census posters.) 
     However, it is the Main Data Bank option that leads the way to 
significant collections of additional census data. Some of the gopher 
categories, such as the statistical briefs, the bulletins, and the We the 
People series, will be found under the Documents and Publications 
option. The listing for Population connects to the same four resources: 
a description, the PopClock, projections, and estimates. Papers from 
the Center for Economic Studies are available from both interfaces. 
     Other sections of the Main Data Bank are specifically designed for 
the Web. The Financial Data for State and Local Government and 
Schools section is set up as HTML documents. Although there is a link 
in the gopher server to this data, it requires a Web browser to view. 
The GeoWEB listing also connects to the same three points, but only 
the Web version offers a connection to the GeoWEB home page. One 
resource  only available on the Web home page is the Tips on Genealogy 
section, which provides information and the beginnings of a FAQ for 
people seeking genealogical information from the Census Bureau. Since 
it is primarily ASCII textual information, it may well be a link from 
the gopher server in the near future. 
     The most promising feature of the WWW Census information lies in 
the Lookup program. One of my criticisms of the gopher server was the 
lack of depth in the statistics available. Few reports covered any 
geographic area smaller than a state; and nowhere did the gopher offer 
detailed population figures, the statistic that first springs to mind at 
mention of the Census Bureau. However, hundreds of CD-ROMs from the 
1990 Census of Population and Housing have been released with user-
friendly software. The experimental Lookup program bridges the gap 
between the Internet and CD-ROM resources in an impressive manner. 
     Lookup requires a WWW browser that supports forms, such as Lynx 
2.3, or any of the current versions of Mosaic. It is available under the 
Data Access Tools option, or directly at either of these URLs: 

http://cedr.lbl.gov/cdrom/lookup  or
http://bigsur.lbl.gov/cdrom/lookup
 
     Data from the census summary tape files 1A, 3A, and 3C are 
available. The Web forms feature is used to select the desired 
variables, such as geographic area and demographic breakdowns. Once 
all the criteria have been selected, Lookup offers the data in one of 
three ways: HTML for direct browsing with a Web client, tab-delimited 
format for use in a spreadsheet, and CODATA format. 
     According to its own files, the Census Bureau is "Factfinder for the 
Nation" and "collects data about the people and economy of the United 
States and produces a wide variety of statistical data products 
including printed reports, statistical briefs, and computer files on 
tape and CD-ROM media." With its multifaceted presence on the 
Internet, the Census Bureau is becoming another major source for 
trustworthy information in cyberspace. And if the Lookup project is 
successful at providing an acceptable interface between its CD-ROM 
products and the net, it may lead the way for an even greater expansion 
of Internet-accessible statistics. 

----------------------------------------------------------------

Communications to the author should be addressed to Greg R. Notess, Montana State University Libraries, Bozeman, MT 59717-0332; 406/994-6563; greg@notess.com ; http://www.notess.com.

Copyright © 1995, Online Inc. All rights reserved.