BRFSS Maps: Visualizing Oral Health Surveillance Data

1
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
2
Outline of Presentation

• Basic concepts data tables vs. maps
• Purpose and functionality of BRFSS Maps
• Future enhancements to BRFSS Maps
• Summary

3
Data Tables vs. Maps

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

4
What to Use Data Table or Map?
5
Multivariate Relationships Tables
6
Multivariate Relationships Map Comparison
7
(No Transcript)
8
(No Transcript)
9
Multivariate Relationships Multivariate Map
10
Maps Static vs. Interactive

• 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• If data are updated - map is outdated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

11
BRFSS Maps

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

12
Purposes

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

13
Functionality

• 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

14
BRFSS Maps Development Timeline

• 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

15
Initial Map Screen
16
Prevalence Data
17
User Control over Mapping Options

• Natural breaks
• Quantiles
• Equal intervals
• Standard deviations

18
Mapping by Standard Deviations
19
Mapping by Quintiles
20
Zooming and Panning
21
Saved Graphic Image
22
Download GIS Data
23
New DevelopmentsSince Initial Release

• Improved internet response time in Version 1.1
• Adapted for rapid response to emerging public
  health issues (e.g., influenza vaccination
  surveillance, 2004-05)

24
System Components

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

25
System Architecture
Get the data
Send map request
I want to see a map!
Web Server
HTTP request
Retrieve dataset
image streamed back
Web Browser
Data being mapped are passed to GIS service
image streamed back
ArcIMS generates map image
GIS Server
26
GIS Web Service

• 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
• Sensitive datasets are not exposed
• Performance enhancement
• Very fast response time
• No lengthy browser downloads

27
BRFSS Maps Future Enhancements

• Version 2.0
• New user interface
• Redesigned map frame
• New visualization tools
• Versions 3.0 and 4.0
• Visualization of multivariate relationships
• Visualization of county-level data

28
New User Interface and Map Frame
29
New Visualization Tools

• Natural breaks
• Quantiles
• Equal intervals
• Standard deviations
• Manual breaks

30
Visualization of Multivariate Relationships

• 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

31
Multivariate Relationships Multivariate Map
32
Mapping County-level BRFSS Data

• 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

33
Raw Data Counties gt 50 interviews

34
Raw Data All Counties
35
Spatially-Smoothed All Counties
36
Isopleth Map Spatially-Smoothed
37
Cognitive Issues

• Rate retrieval and recall
• Pattern recognition and recall
• Map comparison
• Visual dominance
• User preferences

38
Technical Issues

• Data-related
• Cartography-related
• Internet-related

39
Summary

• 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 mappingnot merely an
  out-of-the-box application

40
Questions or Comments?

• Jim Holt
• jgh4_at_cdc.gov
• 770-488-5510
• BRFSS Maps
• http//apps.nccd.cdc.gov/gisbrfss/

41
(No Transcript)
42
Visualization of County Level Data

• Planned for BRFSS Maps version 3.0 or 4.0
• Many technical and cognitive issues