Visualizing Ten Years of Quantitative Color Schemes

Travis White
University of Kansas
#nacis2015

Transcript

1. View Slide

2. (color that represents quantitative data)
   Qc = Quantitative color schemes
   

   View Slide

3. Existing research & resources
   indicate how Qc should be used…
   

   View Slide

4. Existing research & resources
   indicate how Qc should be used…
   How are mapmakers using Qc?
   

   View Slide

5. Qc Map Review & Evaluation
   Key parameters
   ∙2004-2013
   ∙geographic journals from multiple fields
   ∙no cartographic articles
   

   View Slide

6. Qc Map Review & Evaluation
   Key parameters
   ∙2004-2013
   ∙geographic journals from multiple fields
   ∙no cartographic articles
   ∙at least one map selected per article
   ∙univariate maps only
   ∙no elevation maps
   

   View Slide

7. Qc Map Review & Evaluation
   Key parameters
   ∙2004-2013
   ∙geographic journals from multiple fields
   ∙no cartographic articles
   ∙at least one map selected per article
   ∙univariate maps only
   ∙no elevation maps
   440 Qc maps from 315 articles
   

   View Slide

8. Human
   Human
   & Physical
   Physical
   

   View Slide

9. 100%
   80%
   60%
   40%
   20%
   0%
   2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
   Qualitative color scheme rates
   11%
   29%
   42%
   71%
   

   View Slide

10. Question 1
    Which Qc schemes were used?
    How appropriately? By whom?
    

    View Slide

11. Question 1
    Which Qc schemes were used?
    How appropriately? By whom?
    Question 2
    Which colors were used?
    How appropriately? By whom?
    

    View Slide

12. Question 1
    Which Qc schemes were used?
    How appropriately? By whom?
    Question 2
    Which colors were used?
    How appropriately? By whom?
    

    View Slide

13. 2 Qc mapper personas:
    

    View Slide

14. Persona 1
    ∙maps human data
    ∙works w/ unipolar data
    ∙prefers choropleth
    (enum. units)
    ∙uses fewer data classes
    2 Qc mapper personas:
    

    View Slide

15. Persona 1
    ∙maps human data
    ∙works w/ unipolar data
    ∙prefers choropleth
    (enum. units)
    ∙uses fewer data classes
    2 Qc mapper personas:
    Persona 2
    ∙maps physical data
    ∙works equally between
    polarities
    ∙prefers surface data
    ∙uses more data classes
    

    View Slide

16. Useful color assessment requires
    accurate color identification, but…
    consistent accurate sampling
    not possible with unclassed schemes:
    Cloetingh 2007
    , Koch 2009, Fischer 2001, Salvati 2013
    

    View Slide

17. Gimpel 2004, Combes 2005, Wahl 2012, Holand 2013
    However…
    consistent, accurate sampling
    is possible with classed Qc
    

    View Slide

18. Classed Qc sampling process
    Photoshop CC 2015
    sRGB IEC61966-2.1 color profile
    RGB color image mode
    example: Escutia 2005
    Confirm the color
    profile, copy legend
    to a new layer
    Edit out
    pixel noise
    

    View Slide

19. Classed Qc sampling process
    Photoshop CC 2015
    sRGB IEC61966-2.1 color profile
    RGB color image mode
    example: Escutia 2005
    Average
    the edited
    color pixel
    values
    Point sample
    the averaged
    color values
    [eyedropper]
    

    View Slide

20. DISCLAIMER
    Sampling & evaluations based
    on digital versions of maps,
    not original map files.
    Qc scheme appearances are
    inconsistent across display devices.
    

    View Slide

21. Sequential
    Diverging non-spectral
    Spectral
    Traffic
    Other
    Qc Overview
    5 primary schemes
    

    View Slide

22. Sequential Red/Orange/Yellow
    ∙n = 99 (31% of all Qc maps)
    ∙primary scheme for choropleths
    ∙preferred scheme for human data
    

    View Slide

23. Sequential Green/Blue/Purple
    ∙n = 99 (31% of all Qc maps)
    ∙primary scheme for choropleths
    ∙preferred scheme for human data
    

    View Slide

24. Sequential Yellow/Orange/Brown
    ∙n = 99 (31% of all Qc maps)
    ∙primary scheme for choropleths
    ∙preferred scheme for human data
    

    View Slide

25. Sequential Examples
    

    View Slide

26. Sequential Examples
    

    View Slide

27. Sequential Examples
    

    View Slide

28. Diverging non-spectral (critical class)
    ∙n = 86 (27% of all Qc maps)
    ∙common on both choro & surface maps
    ∙more frequently used for physical data
    

    View Slide

29. Diverging non-spectral (critical break)
    ∙n = 86 (27% of all Qc maps)
    ∙common on both choro & surface maps
    ∙more frequently used for physical data
    

    View Slide

30. Diverging non-spectral Examples
    

    View Slide

31. Diverging non-spectral Examples
    

    View Slide

32. Spectral (non-diverging)
    ∙n = 76 (24% of all Qc maps)
    ∙overwhelmingly used on surface maps
    ∙overwhelmingly used for physical data
    

    View Slide

33. Spectral (diverging)
    ∙n = 76 (24% of all Qc maps)
    ∙overwhelmingly used on surface maps
    ∙overwhelmingly used for physical data
    

    View Slide

34. Spectral Example
    

    View Slide

35. Spectral Example
    

    View Slide

36. Spectral Example
    

    View Slide

37. Traffic (diverging & non-diverging)
    ∙n = 26 (8% of all Qc maps)
    ∙more common on surface maps
    ∙used for half of all human + physical data
    

    View Slide

38. Traffic Example
    

    View Slide

39. Other (hypso, thermal, undefined…)
    ∙n = 35 (11% of all Qc maps)
    ∙no noteworthy patterns
    

    View Slide

40. Other Example
    

    View Slide

41. Other Example
    

    View Slide

42. Accentuating the positive
    ∙Sequential & Diverging schemes used
    appropriately
    ∙Qc schemes rarely lost in visual hierarchy
    ∙Color deficiency not too problematic
    (ignoring Spectral & Traffic schemes)
    

    View Slide

43. Accentuating the negative
    ∙Too many classes!
    ∙Poor color grading
    (cannot distinguish colors within a scheme)
    ∙Poor matching btwn data type & Qc
    (e.g., unipolar data + diverging scheme)
    ∙Illogical sequencing of colors
    ∙Image quality matters!
    

    View Slide

44. Recommendations
    ∙Educate & Inform!
    mappers need instruction and guidance
    ∙Improve graphics publication guidelines
    ∙Improve or limit software/package offerings
    ∙A Brewer for non-choro mapping (?)
    

    View Slide

45. “GOOD ENOUGH”
    is not good enough!
    

    View Slide

46. Appropriate Qc usage
    ≠
    GOOD DESIGN
    

    View Slide

47. Thank You!
    Expanded slides & RGB profiles available at:
    References
    Bordi, I. and A. Sutera. 2004. Drought variability and its climatic implications, Global and Planetary Change 40, 115–127
    .
    Brewer, Cynthia A., 2013. http://www.colorbrewer2.org, accessed October 4, 2015
    Gillespie, T, Foody, G., Saatchi, S. 2008. Measuring and modelling biodiversity from space, Progress in Physical Geography 32, 203.
    Groisman, P
    . Y. 2007
    . Potential forest fire danger over Northern Eurasia: Changes during the 20th century, Global and Planetary Change, 56, 371–386.
    Kaplan, D. and G. Sommers. 2009. An Analysis of the Relationship Between Housing Foreclosures, Lending Practices, and Neighborhood Ecology: Evidence from a Distressed
    County, The Professional Geographer, 61:1, 101-120.
    Lewis, L. 2010. Biogeography and Genetic Diversity of Pearl Millet (Pennisetum glaucum) from Sahelian Africa , The Professional Geographer, 62:3, 377-394.
    Lombard, A. 2005. Contribution of thermal expansion to present-day sea-level change revisited, Global and Planetary Change 47
    , 1–16.
    Maliszewski, P
    . and M. Horner. 2010. A Spatial Modeling Framework for Siting Critical Supply Infrastructures , The Professional Geographer, 62:3, 426-441.
    Morehouse, B and S. O’Brien. 2008. Facilitating Public Involvement in Strategic Planning for Wildland Fire Management , The Professional Geographer, 60:4, 495-507
    .
    Setton, E. 2010. Gender Differences in Chronic Exposure to Traffic-Related Air Pollution—A Simulation Study of Working Females and Males, The Professional Geographer, 62:1, 66-
    83.
    Taus, A. 2013. Conversion to Organic Farming in the Continental United States: A Geographically Weighted Regression Analysis, The Professional Geographer, 65:1, 87-102.
    Wen, T., Lin, M., Fang, C. 2012. Population Movement and Vector-Borne Disease Transmission: Differentiating Spatial–Temporal Diffusion Patterns of Commuting and Noncommuting
    Dengue Cases, Annals of the Association of American Geographers, 102:5, 1026-1037
    .
    

    View Slide