Visual Display and Analysis of Geo-referenced Data

ABSTRACT:
Many characteristics of people and their environment vary by location. Summary statistics that ignore this variation can mask important and informative differences by place. Location-specific (geo-referenced) data are usually displayed on a thematic map, often with areas shaded by color depicting different levels of the mapped variable. However, a poorly designed map can mislead the viewer and lead to erroneous conclusions about patterns in the underlying data. Cartographic guidelines that exist for navigational maps are not necessarily helpful for designing data maps. In this talk, we illustrate early US county-level data map design using cancer rates, examining how the maps led to important public health findings. We then summarize cognitive research that led to improvements in thematic map design for both static display and exploratory data analysis. We then will demonstrate visual data exploration by means of the linked micromap plot (LM) which displays a series of small maps and graphs of the same data, linked by color. This provides a visual analysis of the geographic patterns in the data and associations among multiple geo-referenced variables. Together, the mapping tools illustrated in this talk can be used for initial display, exploration and communication of geo-referenced data.

BIO:
Linda Williams Pickle, Ph.D.
Principal and Chief Statistician
StatNet Consulting, LLC

Dr. Pickle received her Ph.D. in Biostatistics from the Johns Hopkins Bloomberg School of Public Health and is an elected Fellow in the American Statistical Association. She has over 40 years of experience developing better statistical methods and data visualization tools for analyzing and presenting health-related data, particularly by geographic area. While at the National Center for Health Statistics and the National Cancer Institute, she published over 150 articles in the medical and statistical literature and 3 major atlases showing the geographic patterns of disease rates. She published Visualizing Data Patterns with Micromaps with Daniel B. Carr in 2010, a book describing methods to link graphs and maps so that geographic and attribute patterns can be examined simultaneously. Dr. Pickle coordinated GIS activities at NCI during 1999-2007 and her research in spatial statistical models led to improved techniques for modeling incidence rates at the U.S. county level, a method now used by the American Cancer Society to predict the number of new cancer cases in the current year for their Cancer Facts and Figures annual publication.