Building Data-rich Web Sites: The Integration Projects of the Minnesota Population Center William C. Block IASSIST 2006 Ann Arbor, Michigan, USA 24 May.

Building Data-rich Web Sites: The Integration Projects of the Minnesota Population Center

William C. BlockIASSIST 2006

Ann Arbor, Michigan, USA 24 May 2006

Overview

• Introduction to the problem (Powerpoint)– Rich data– Resulting Design challenges

• A look at three MPC data dissemination sites and issues related to their design

• Conclude with a summary of the design issues (return to Powerpoint)

Caveat: Using a broad definition of Design in this session

• Session Title: “Effective Design for Data-Rich Websites”

• Design can be more than just front-end, User Interface issues

• Back-end design is certainly relevant to building a data-rich website

• Even more abstract: Effective Design during the whole project process can help achieve broader project goals

Problem: A (growing and changing) mountain of data to disseminate

• IPUMS-USA– Began as just historical census data– grew into IPUMS-USA– Integrated U.S. Census and American Community

Survey data., 1850-2004

• IPUMS International– Currently 19 census samples/8 countries– Funded to do 150 samples/50 or so countries– ~1 billion person records when complete

Integrated Health Interview Survey

• Integrated National Health Interview Survey data, 1963-2004

• ~20,000 variables, including (but not limited to): socio-demographics, general health status, injuries, chronic conditions, health insurance coverage, access to and utilization of medical, dental, and hospital services, health promotion activities, immunizations, AIDS knowledge and attitudes, disability, alcohol and tobacco use, and cancer control.

• Complex file structure (very different than hierarchical census data!)

Other Rich Datasets of the MPC

• Aggregate Census Data (NHGIS)

• National Survey of Families and Households (NSFH) – prototype as part of DSDR project– Multiple wave survey– Complex file structure

• American Time Use Survey

• High School Exit Exams

In brief…

• A mountain of data…much of it very different and more complex than our earlier data.

• How do you design an effective website for such a data-rich environment?

• Can you even build one web infrastructure to accommodate such disparate and sizeable data?

A look at three websites…

• IPUMS-USA (old microdata design)

• IPUMS International (new microdata design)

• IHIS (first foray into non-census data; New microdata Design)

Summary of the Design Issues Facing MPC

• Improved tools for navigating so much information

• Ability to add different kinds of data (such as Phase I of the health data)

• Flexibility to add modules for radically different kinds of data (such as Phase II of IHIS)

• Ability to serve multiple audiences with one website– The public users of our data– The researchers doing the integration

Conclusion:

• Effective Design of Data-rich websites encompasses much more than just front-end concerns. Effective design should permeate the entire project process from beginning to end, from back-end to front-end, and help you achieve broader project goals for data-rich websites.

Thank you!

• Questions?

• Comments?

• Criticisms?

• [email protected]