IBM/FactFinder - Automatically Producing Layouts from Content

Customer

"I'm delighted with Fenestra's R&D efforts in creating timelip, robust, and practical solutions. They looked at our problems in a unique way and streamlined efficiencies using existing technologies." - Brenda Goodwin Dixon, IBM Global Services

In 1997 the Census Bureau contracted with IBM to develop, implement and run American FactFinder.

American FactFinder (previously known as the Data Access and Dissemination System, a/k/a/ DADS) provides Census products to meet a variety of data needs for different segments of the data user community. Various media are created including CD-ROM, Internet, DVD and printed reports.

Opportunity

The Decennial Census survey, taken in 2000 by the U.S. Census Bureau, yields a wealth of data about 115.9 million housing units and 281.4 million people across the United States. IBM was searching for an efficient way to streamline the publishing of data in the form of web pages and books.

Desires of Key Users

Fenestra was retained to help prepare printed reports. More than 1,000 tables and 52 books summarizing 2000 Decennial Census data will be generated in this project.

The books are printed to exacting specifications with many specialized line breaking and pagination rules.

Solution

The GIDS Autoformatter technology (see the U.S. Census Bureau case study), designed for use with the 2002 Economic Census, automates the process of creating rules-driven reports. The work was broken into two phases: prototype/discovery and production.

Technologies Used

The GIDS Autoformatter includes a sophisticated paper and electronic typography system that can be used to automatically produce layouts from content information. For American FactFinder, we used this typography system to produce data dissemination reports.

In both paper and electronic methods, the information contained in the report (what we call "content") was presented to our software in XML format, and the layout was automatically produced using rules and templates.

Outcome

The exact format of the rules was determined by a team effort between Census, IBM and Fenestra. Following the completion of the prototype phase, 52 states books (including DC and Puerto Rico) with an average of 26 tables per state, were generated in only two days. The resulting printed output comprised approximately 26,000 pages.

Fenestra/Client Relationship

It is expected that the Decennial Census will continue to use this technology for future publications of complex and lengthy surveys, needing only to slightly alter it for new fields, based on changing times. The Economic Directorate may also leverage this technology for dissemination of the 2002 Economic Census data.

Fenestra/Client Relationship

IBM and Census are satisfied with the performance of this technology, and plan to review other opportunities for its implementation in other projects.