Maps

6. Maps#

6.1. Map#

  1. Definition: a map is a collection of spatially defined objects (Mark Monmonier)

  2. Beyond mapping

    • Map as analysis vs Map as presentation

      • Geo-visualization

      • Geospatial visual analytics

    • Exploratory spatial data analysis (ESDA) - Luc Anselin

      • Spatial regimes: Spatial regimes are a form of spatial heterogeneity, which implies structural differences across space.

      • When a variable is characterized by distinct distributions (e.g., with a different mean or variance) for different geographic subregions, these subregions might point to the existence of spatial regimes.

6.2. Traditional Knowledge Discovery#

  1. Deductive approach

    • Hypothesis first, data later

  2. Inductive approach

    • Data first, hypothesis later

  3. Abductive approach

    • Pattern discovered along with hypothesis

    • Interaction between data exploration and human perception

6.3. Map Design Primer#

6.4. How to Lie with Maps#

  1. Manipulate map design parameters

    • Scale, Symbols, Legends, Colors, Intervals

  2. Choice of Projection

    • Larger areas seems more important

    • Conformal = Preserve angle

    • Equal area = Preserve area

    • Equal distant = Preserve distance

    • Azimuthal = preserve direction

  3. Human Perception can be tricked

6.5. Choropleth Map#

  1. Visualizing a spatial distribution

    • Natural Breaks VS Quantile

    • Natural Breaks use clustering algorithm (minimum the heterogeneity within classes)

    • Natural Breaks have different number of observations per category

6.6. Continuous Statistical Maps#

6.7. Percentile Map#

  1. Special form of Quantile Map - Percentiles

  2. 6 categories instead of 100 categories

    • < 1%, 1-10%, 10-50%, 50-90%, 90-99%, >99%

  3. Emphasis on Extremes - Away from median

  4. Only works well for large data sets

    image.png

6.8. Box map plot (Luc Anselin)#

  1. Box and whiskers plot

    • Identify shape of distribution and outliers

    • Focus on median

  2. Inter quartile range (IQR)

    • Range from 25% to 75%

    • Fence = 75%/25% \(+/-\) 1.5 IQR or \(+/-\) 3 IQR

      • Outliers = outside the fence

    image.png

6.9. Standard Deviation Map#

  1. Based on standardized data value

    • Mean = 0, standard deviation = 1

  2. Intervals correspond to one standard deviation

  3. Outliers are more than 2 standard deviations from the mean

    img

6.10. Categorial Statistical Maps#

6.11. Co-Location Map#

  1. Unique value map or Categorical map

    • For discrete categories

  2. Map overlay

    • Map algebra

    • Matching categories between two or more maps

    img6

  3. Multivariate categorical association

    1. Transfer box plot into categorical map (1-6)

    2. Find the overlap of the categories (rank)

    img7

6.12. Cartogram#

  1. Areal unit proportional to variable of interest

  2. Avoid misleading effect of area

    img8

  3. Use transformed shapes

    • Circular cartogram and Contiguous cartogram

    img9

6.13. Conditional Map#

  1. Special case of trellis/facet/conditional graphs

  2. Micro-map matrix

    • Conditioning variables on axes

    • Matrix of mini maps for the variable of interest conditioned by values on the axes

    img10

Contents