Storm Uncertainty Representation: User Responses to Novel Visualization

Jennifer Smith Mason – Penn State University
Jason Dykes, Jo Wood, & Aidan Slingsby – City University, London

Transcript

1. Storm Uncertainty Representation: User Responses to Novel Visualization Jennifer Smith Mason – Penn State University Jason Dykes, Jo Wood, & Aidan Slingsby – City University, London NACIS October 11, 2013
2. Outline Why Visualize? Problems Design criteria Solutions
3. Why hurricanes?
4. Why hurricanes? Spatial Temporal Plan
5. Why visualize hurricanes? !  Public use !  Context !  Concern !  Evacuation !  Efficiency
6. Why visualize hurricanes? Informa(on  Overload!  
7. Problems with Existing Visuals
8. "Hurricane-warning graphics, despite admirable attempts by the forecast community to make user-friendly products and to reach out to emergency managers and the media, are subject to misinterpretation by many members of the public." (Broad et al., 2007, p. 665)
9. Problems with Existing Visuals
10. Problems with Existing Visuals ✔
11. Qualitative Feedback Cone: !  Want better way to see dates !  Want wind speeds Wind Speeds: !  Mark trajectory !  Looked realistic
12. Design Constraints !  Dynamic nature ! Salient !  Wind speeds – multivariate (one map) !  Central track uncertainty !  Time
13. Research Questions !  Which visualization? !  Accurately rank locations? !  Confident?
14. Data http://www.nhc.noaa.gov/gis/ •  Forecast track •  Cone of uncertainty •  Wind field •  Wind speed probabilities (39, 58, 74 mph)
15. Possible Solutions Noise Annotation Lines Points
16. Possible Solutions
17. Transparency = Probability
18. Each cell square divided into 3 horizontal strips Probability: transparency
19. Height (horizontal stripes): proportion of probabilities Transparency: sum of them.
21. Thank You!