BRFSS Maps Visualizing Oral Health Surveillance Data James

BRFSS Maps: Visualizing Oral Health Surveillance Data James B. Holt, M. P. A. , Ph. D. Centers for Disease Control and Prevention Pittsburgh, PA May 4, 2005

Outline of Presentation n n Basic concepts: data tables vs. maps Purpose and functionality of BRFSS Maps Future enhancements to BRFSS Maps Summary

Data Tables vs. Maps n n Different ways of looking at the data Specific values (tables) or spatial patterns in the data (maps)? • In other words: “detail” vs. “simplification” n n Data for analysis (tables) or data for visualization (maps)? Access to ancillary information (tables)? • Confidence intervals • Cell sizes n Comparison of two or more variables

What to Use: Data Table or Map?

Multivariate Relationships: Tables

Multivariate Relationships: Map Comparison

Multivariate Relationships: Multivariate Map

Maps: Static vs. Interactive n n Static = one-of-a kind snapshot • If data are updated - map is outdated • Okay if you anticipate a one-time need • Can be produced in a graphics package or a geographic information system (GIS) Interactive = dynamic • If data are updated – map automatically updates • Good for one-time needs as well as periodic updates • Can be produced via a GIS, with internet-enabled interactivity, controlled by end-user

BRFSS Maps n n Interactive GIS-enabled Internet-based mapping application Visual depiction of BRFSS data • States • Metropolitan and Micropolitan Statistical Areas (MMSA)

Purposes n Visualization and exploratory spatial data analysis • Highlight regional patterns and interstate variations • Prompt interesting research questions n n BRFSS data in a GIS-friendly format Encourage the use of geographicallyreferenced risk factor data

Functionality n n n n View state and MMSA data together or separately User selectable class breaks Zoom, pan, and prevalence rate retrieval User selectable labeling Image downloading GIS data downloading Access to mapping guidance

BRFSS Maps Development Timeline n n n n Conceptual development: July 2003 Technical planning & layout: October 2003 User feedback: February 2004 Version 1. 0: August 2004 Enhancement planning: continuous Version 1. 1: January 2005 Version 2. 0: June 2005, expected

Initial Map Screen

Prevalence Data

User Control over Mapping Options n n Natural breaks Quantiles Equal intervals Standard deviations

Mapping by Standard Deviations

Mapping by Quintiles

Zooming and Panning

Saved Graphic Image

Download GIS Data

New Developments Since Initial Release n n Improved internet response time in Version 1. 1 Adapted for rapid response to emerging public health issues (e. g. , influenza vaccination surveillance, 2004 -05)

System Components n n n Web server • IIS (Microsoft web server software) • ASP. Net (webpage development software) Database server • SQL Server 2000 n NCCDPHP data repository GIS server • Arc. IMS n Internet GIS development n map generation

System Architecture I want to see a map! Web Server HTTP request Get the Send data… map request Database Server Retrieve dataset image streamed back Web Browser image streamed back Data being mapped are passed to GIS service Arc. IMS generates map image GIS Server

GIS Web Service n n Mapping capability easily integrated into existing Web sites Increased data security • GIS service does not require database access Mapped data sent with each map request n Sensitive datasets are not exposed n n Performance enhancement • Very fast response time • No lengthy browser downloads

BRFSS Maps Future Enhancements n Version 2. 0 • New user interface • Redesigned map frame • New visualization tools n Versions 3. 0 and 4. 0 • Visualization of multivariate relationships • Visualization of county-level data

New User Interface and Map Frame

New Visualization Tools n n n Natural breaks Quantiles Equal intervals Standard deviations Manual breaks

Visualization of Multivariate Relationships n n n n Many options… Map-to-map comparison Bivariate map (e. g. , Census map) Proportional symbol over choropleth Stacked bars over choropleth Conditioned choropleth Would require cognitive testing and technical feasibility study

Multivariate Relationships: Multivariate Map

Mapping County-level BRFSS Data n n Data availability (sample size) Statistical noise Spatial smoothing & small area ests. Purpose and limitations of maps of spatially smoothed sample data • Pattern recognition • Rate retrieval

Raw Data *Counties > 50 interviews

Raw Data All Counties

Spatially-Smoothed All Counties

Isopleth Map Spatially-Smoothed

Cognitive Issues n n n Rate retrieval and recall Pattern recognition and recall Map comparison Visual dominance User preferences

Technical Issues n n n Data-related Cartography-related Internet-related

Summary n n BRFSS Maps enables visualization of oral health and other public health surveillance data Development emphasized usability and utility Future developments will focus on visualizing multivariate relationships, county-level mapping and additional functionality Innovation in mapping…not merely an “outof-the-box” application