Statistical Resources

Meta Sites
These sites are collections of links to data resources on the Web. They all have their own quirks in selection of resources, organization, annotation and site-searching capabilities.

Larry Shankman's Statistical Page

Shankman, a librarian at Mansfield State University, has put together an excellent guide to statistical sources. This site is extremely well-organized for reference. There are eight general subject areas. Within each area, Shankman lists standard print sources (limited to the reference collection at Mansfield) with LC call numbers, all annotated generously, and provides links to descriptions of locally-held CD-ROM products and their online equivalents. He does some brief comparisons, but not in detail. There is no overall index or any way to search the site for sources by subject or table data descriptions.

University of Michigan Documents Center

Another excellent site for online resoures. Does not list or compare CDs or print resources. Frames index allows easy searching for fairly specific subject areas covered on the site. Especially good for online international statistics produced both by organizations like the UN and individual countries. Resources are annotated.

Census Bureau

The Bureau site itself is well done. Beyond the home page, the "Resources" page has a complete index of topics (more complete than, say, Michigan's frame index), as well as an initial letter locator for the index, and a search engine (TOPIC from Verity, Inc., one of the better ones) which is underemployed searching the site's 136 pages at a very high level, e.g. "agriculture". Note, too, that the results of the search on "agriculture" and the results from choosing "agriculture" from the index are very different. The topical pages themselves are highly recommended for basic information, including definitions, descriptions of print and CD-ROM resources, and links to related sites. This "topic-centered" approach can be very helpful in leading to some important, but unfamiliar, or unthought-of, resources.

CenStats is a subscriber-only product from the bureau (and Cornell subscribes) which allows access to some resources with non-PDF formats. For example: County Business Patterns, USA Counties, and some unique resources, such as Building Permits. Output is text on a Web page. Occasionally this makes the re-use of the data difficult.

FedStats

FedStats provides the "rest of the story"--access to data from other government agencies than the Bureau of the Census (although it includes their products also). The TOPIC search engine is put to better use at this site, which is indexed down to the variable level for many tables. Example: a search on "filberts" brought up 11 sites. An excellent site for specialized statistics online that aren't indexed or linked to from other sites. Output options vary.

StatUSA

STAT-USA gathers business, economic, and trade data from over 40 government agencies, consolidates this information, and then provides it over a range of electronic formats. It provides a fairly limited range of products, the most notable being National Trade Data Bank, but does have a lot of business-oriented data not available from any other site. StatUSA is divided into two sections for international trade. Beyond the data offered from these pages, researchers can move into the State of the Nation Library and the International Trade Library for even more arcane data sets. Both sites have some search capability and well-organized lists, but some of the information is still difficult to find. Output options vary.

CIESIN and SEDAC

Amazing DataViewer--does much that the Census Bureau's EXTRACT does and more! DDViewer provides rapid data mapping, viewing and analysis to the internet community. There are Java and non-Java versions. Latest: Java 3.0. The primary data set is the US Census of Population and Housing, 1990, but there are specialized environmental data sets also. Spectacular map-building, but data can also be presented as text for cutting and pasting into a spreadsheet.
The advantages over EXTRACT and GO on CD-ROM are these: You don't need to load separate CDs for each state, or choose and redo your choice of data items from different "Catalogs". Data is now down to the tract level, so that it's possible to choose, for example, the border tracts of Pennsylvania and New York Counties together and choose data items from widely different groups (population, education, housing). With the CD-ROM this would take two different CDs and several different extractions from different catalogs on those CDs, then many merges of data to form a single table. The disadvantages: no data by block group or blocks; limited number of data items; fewer output options, some geographies, such as Metropolitan Areas, Congressional Districts, and Native American Reservations aren't covered.

Government Information Sharing Project at Oregon

Excellently presented access to a select group of primary government statistical sets: (USA Counties, REIS, Economic Censuses. Many sites actually link to Oregon for their access. Oregon has a consistent interface which allows one to choose the geography from an image map and choose variables from a list. Output is monospaced text on an Web page.

WebEc

One of the best collections of links for economic data and international statistics.

San Diego

An excellent search engine which searches a very complete and exhaustively annotated list of data sources. Now includes an experimental "data on the web" search. Very complete and uncomplicated.

Virginia

Complements Oregon with many interactive data sites, including REIS and CCDB.