QCBS R workshop - Visualization

Transcript

1. Plotting in
   Open Source Statistical
   Programming
   map courtesy of James Cheshire (spatialanalysis.co.uk/2012/02/great-maps-ggplot2/)
   

   View Slide

2. Tables
   Data
   Graphs
   Statistics
   Understanding
   Sigmaplot
   Excel
   SAS
   Why do your plotting in R?
   Simpler workflow:
   

   View Slide

3. Tables
   Data
   Statistics
   Understanding
   Graphs
   Why do your plotting in R?
   Simpler workflow:
   

   View Slide

4. Why do your plotting in R?
   Flexible and beautiful graphics:
   

   View Slide

5. Why do your plotting in R?
   Flexible and beautiful graphics:
   

   View Slide

6. Why do your plotting in R?
   Flexible and beautiful graphics:
   

   View Slide

7. Enough slides! Now for a
   live demo
   

   View Slide

8. *
   Challenge 1:
   Building a plot layer by
   layer
   1. Use the 'iris' dataset to build the following three
   plots:
   a. A plot of Sepal.Length vs. Sepal.Width
   b. The same plot, but coloured by 'Species'
   c. A plot of Petal.Length vs. Species
   

   View Slide

9. *
   Challenge 2:
   Playing with aesthetics
   and geoms
   1. Use the qplot function and the 'iris' dataset to re-
   build the following three plots:
   a. A plot of Sepal.Length vs. Sepal.Width
   b. The same plot, but coloured by 'Species'
   c. A plot of Petal.Length vs. Species
   2. For the plot in part 'c', try changing the geom
   argument to "boxplot". Try building a graph with
   both points and boxplot on them
   

   View Slide

10. *
    Challenge 3:
    Faceting and scaling
    1. Try to build the following plots using the iris data:
    a. Petal.Width vs. Petal.Height, faceted in a row
    by Species; and the same plot by column
    b. Sepal.Width vs. Sepal.Height, with the x-axis
    log-transformed, the y-axis log-transformed,
    and both axes log-transformed
    

    View Slide

11. ggplot2: using the Grammar of
    Graphics
    ggplot2: a different way of plotting: building
    plots by describing what you want, and let it
    figure out how to plot it.
    

    View Slide

12. ggplot2: using the Grammar of
    Graphics
    ggplot2:
    Builds plots out of simpler components:
    ● data
    ● aesthetics: e.g. coordinates, colour, size of points
    ● geoms: the type of plot layer: line, point, etc.
    ● facets: organizing graphs into subplots
    ● transforms: logs, square roots, continuous to discrete
    etc.
    ● stats: functions of the raw data
    ● coordinates: e.g. Euclidean, map coords, polar
    ● themes: font size, grids, etc.
    

    View Slide

13. Further reading
    Plotting in general:
    R Graphics
    The R Book
    ggplot2 references:
    The creator, Hadley Wickham's website
    The in-depth documentation website
    The book: ggplot2: Elegant Graphics for Data
    Analysis
    

    View Slide