U.S. Census Bureau

Guide to the Economic Census
Guide to the
2002 Economic Census


2002 Economic Census on American FactFinder®

American Factfinder

American FactFinder (AFF) is a database system that gives you access to the same Economic Census data that are also published in PDF. The advantage with FactFinder is that you can search directly for the data you want, can display data for all available sectors at once, and can download the results for further manipulation.  AFF also provides access to all of the data published on DVD-ROM.



Limitations of AFF

American Factfinder

Scope

AFF includes all 2002 Economic Census geographic area series, industry series, and subject reports. AFF does not certain data sets released only in html format--including advance and preliminary data subsequently superseded by other files.

Coverage of the 1997 Economic Census in AFF was more limited, omitting ZIP Code Statistics, Nonemployer Statistics, and all of the 1997 Economic Census data sets that are classified by SIC.

Limited number of records and geographic areas

AFFdoes not allow the user to manipulate or download more than 15,000 records at a time. During data retrieval in large files, the user must "Filter Rows", reducing the number of areas and/or industries viewed, until the 15,000 maximum is observed. Another limitation is that no more than 3,500 geographic areas can be retrieved at any one time. That allows retrieval of all counties nationwide, but not all cities (of which there are over 10,000 in the Economic Census).

If you need to download data sets with more than 15,000 rows or covering more than 3,500 geographic areas, and cannot conveniently segment the query into multiple retrievals, see the discussions below of the DVD-ROM alternative and FTP downloading.

Quirky download formats

AFF produces download files in comma-delimited or tab-delimited formats. These formats would normally be ideal since comma-delimited files can be imported by nearly all types of data manipulation software. Unfortunately, AFF exports numeric values as if they were character strings, enclosed in quotes, with embedded commas between hundreds and thousands, between thousands and millions, etc. Further, title and other heading lines are included which ignore the field structure of the following data.

The one package that seems to deal well with AFF download files is MS Excel, which treats numbers as values that can be calculated even though they were originally read in as character strings. In some cases it may be advantageous to use Excel as a preprocessor to create .dbf or other output formats for other software.

A disadvantage of using Excel to manipulate output is that it drops leading zeros in FIPS state, county, or other geographic area codes; for example, state code 06 becomes 6. This would be significant to a mapping program. (To reinstate leading zeroes in Excel, right-click on the appropriate column heading, select Format Cells, then select Custom under Category in the Number tab, and then enter into the "Type" box a string of zeroes equal to the number of characters in the code--2 for a state code, 3 for a county code, etc.)

A few packages actually do better importing displays from AFF that have been saved as HTML files through your browsers File...Save As... functionality. Nonetheless, AFF displays to the screen are limited to 100 lines (except for Geography Quick Reports), limiting the usefulness of this technique.

The DVD-ROM alternative

2002 Economic Census DVD-ROMs include all final data sets. Their Windows-based software does not limit the number of records that can be accessed or manipulated. The DVD-ROM software also allows downloading in a wider variety of formats.

Downloading Data via FTP

Another option to get around limitations in the number of records or geographic areas that AFF can export is to download entire data sets, with no filtering, from the Census Bureau's FTP server at www2.census.gov/econ2002/EC/. Files are organized by sector. To determine which file you need, load the data you want interactively in American FactFinder and look for a string like &-ds_name=EC0231A1 as part of the url. The sector number follows EC02 in that string, so the sector would be 31 in this case, and this particular file can be downloaded in entirely from www2.census.gov/econ2002/EC/Sector31/Data/ec0231a1.dat.

Although the FTP files all have the extension .dat, they are in fact valid comma-delimited files, not subject to the quirks noted above for AFF-created .csv files. The files contain no descriptive column headings or industry titles.

Many of these files are quite large and should not be attempted unless a high speed connection is available. The ec0231a1.dat in the example is 17 megabytes. The Economy-Wide Key Statistics File (www2.census.gov/econ2002/EC/Sector00/Data/ec0200a1.dat) is 294 megabytes in size.

The easiest way to access record layout information is load the corresponding data set in AFF, hover over "Options" on the menu bar, and click on "Select Columns". Put a check mark in every box in the "Value" column, and delete every check mark in the "Meaning" column", then click the "Update" button. The resulting display in AFF will correspond directly to the layout of the .dat file, with one exception: each data value is represented in the .dat file with two columns. The first is the numeric value, the second is a flag field for that data item, such as "D" for disclosure or "r" for revised.