Geographic Information Systems and Science for Cancer Control

Geovisualization Tools Development

As part of a broader program of data visualization research, NCI staff and their collaborators have developed a number of methods to better communicate cancer statistics visually, such as by maps and new graph designs. Examples of work in this area are:


Weighted Nonparametric Smoothing (Headbanging)

"Head-banging" is a weighted two-dimensional median-based smoothing algorithm, developed to reveal underlying geographic patterns in data where the values to be smoothed do not have equal variances. The Head-Bang software helps users apply this algorithm to data.

Proportion of residents with no health insurance: CDC Behavioral Risk Factor Surveillance System

  • U.S. map shows the relative proportion of residents without health insurance. The original data contains many small regions.

    Original Data

  • U.S. map shows the relative proportion of residents without health insurance. After head-bang smoothing is applied, regions are larger and more broadly defined.

    Smoothed Data

Data: Pickle & Su, American Journal of Preventive Medicine, 2002
Headbanging: Mungiole et al., Statistics in Medicine, 1999

Return to Top


Tools implemented as ArcGIS Extensions

Return to Top


Conditioned choropleth maps

  • Decomposes a single choropleth map into strata according to 1 or 2 other factors
  • Interactive tool - users control classifications of mapped variable (red/gray/blue) & stratifiers by sliders
Conditioned chloropleth maps

Carr, Wallin, and Carr, Statistics in Medicine 2000 [D]

Return to Top


Sample linked micromap plot of white female cancer mortality rates from 1970 - 1994. Color-coding matches each US state to its rate.

Carr, Wallin, and Carr, Statistics in Medicine 2000

Linked Micromap Plot

  • Color links statistical graphs & maps
  • Micromaps show spatial patterns
  • Supports different glyphs & sorting

Return to Top


Combining Statistical Clustering Algorithms with New Visualization Tools

Screenshot of the Exploratory Spatio-Temporal Analysis Tool

Exploratory Spatio-Temporal Analysis Tool (ESTAT):

  • Time trend plots of cancer rates
  • Parallel coordinate plot of risk factors, confounders
  • Scatter plots, maps
  • Statistical summaries
  • Supports linking, brushing, variable & option selection
  • Screenshot of a time-series plot in the Exploratory Spatio-Temporal Analysis Tool
  • Screenshot of a parallel coordinate plot in the Exploratory Spatio-Temporal Analysis Tool

Alan MacEachren, Penn State University (www.geovista.psu.edu)

Return to Top