第 2 章 Geometries and Aesthetics

Available geometries
Two steps:
- Step 1: What geometries do you see?
- Step 2: With a given geometry, what aesthetics do you observe?

Geometries?
Aesthetics?

2.1 An Example

Geom:
geom_point:
- x, y：varies with observations
- color (線框顏色): constant
- stroke (線框粗細): constant
- fill (內部塗色): constant
- size (點的大小): constant
geom_text:
- x, y: varies with observations
- label: varies with observations
- hjust (文字水平錨點）：between 0 and 1
- vjust (文字垂直錨點）: between 0 and 1

# take Colombia, Normway, and Gemany as an example
data4plot <- list()
data4plot$point$data <- {
  data.frame(
    x=c(52.5, 56, 58.5),
    y=c(48.6, 54.9, 58)
  )
}
data4plot$text$data <- {
  data.frame(
    x=c(52.5, 56, 58.5),
    y=c(48.6, 54.9, 58),
    label=c("Colombia", "Norway", "Germany")
  )
}

myplot <- {
  ggplot()+
    geom_point(
      data=data4plot$point$data,
      mapping=aes(
        x=x, y=y
      ),
      fill="blue", 
      shape=21,  
      color="black",
      stroke=1,
      size=1 
    )+
    geom_text(
      data=data4plot$text$data,
      mapping=aes(
        x=x, y=y, label=label
      ),
      hjust = 1, 
      size = 1 
    )
}
myplot

use econDV2::ggdash() to quickly create a shiny app of your ggplot to fine tune your graph.

2.1.1 econDV2::ggdash

myplot <- {
  ggplot()+
    geom_point(
      data=data4plot$point$data,
      mapping=aes(
        x=x, y=y
      ),
      fill="blue", #input$fill_point
      shape=21, # default point is a circle without border, you need another type of point
      color="black", #input$color_point
      stroke=1, #input$stroke_point
      size=1 #input$size_point
    )+
    geom_text(
      data=data4plot$text$data,
      mapping=aes(
        x=x, y=y, label=label
      ),
      hjust = 1, #input$hjust_text
      size = 1 #input$size_text
    )
}
myplot

myplot <- {
  ggplot() +
    geom_point(
      data = data4plot$point$data,
      mapping = aes(
        x = x, y = y
      ),
      fill = "#2fc1d3",
      shape = 21,
      color = "black",
      stroke = 1,
      size = 12
    ) +
    geom_text(
      data = data4plot$text$data,
      mapping = aes(
        x = x, y = y, label = label
      ),
      hjust = 0,
      size = 12
    )
}
myplot

2.1.2 Script saving

If you are happy with your graph, you may want to make it reproducible and easy to use.

Think of how you want to use it next time.

Turn your script into a function, and save in a R script.

Select the code lines that you want to turn into a function
Code > Extract Function

The result after proper revision:

myGoodPlot <- function(data4plot) {
  myplot <- {
    ggplot() +
      geom_point(
        data = data4plot$point$data,
        mapping = aes(
          x = x, y = y
        ),
        fill = "#2fc1d3",
        shape = 21,
        color = "black",
        stroke = 1,
        size = 12
      ) +
      geom_text(
        data = data4plot$text$data,
        mapping = aes(
          x = x, y = y, label = label
        ),
        hjust = 0,
        size = 12
      )
  }
  myplot
}

To reuse:

source("support/ggplots.R") # suppose this is where your script file is

data4plot <- list()
data4plot$point$data <- {
  data.frame(
    x=c(52.5, 56, 58.5),
    y=c(48.6, 54.9, 58)
  )
}
data4plot$text$data <- {
  data.frame(
    x=c(52.5, 56, 58.5),
    y=c(48.6, 54.9, 58),
    label=c("Colombia", "Norway", "Germany")
  )
}

myGoodPlot(data4plot)

If you replace data4plot with the real data, you get the accurate graph that you want.

2.1.3 Graph output

Saving is tricky. You need to find the viewport ratio. (Different computers are different)
Given your graph width and height, you can save your output as:

dir.create("works")
ggsave(
  filename = "works/week1.png",
  plot=myGoodPlot(data4plot),
  width=930,
  height=400, 
  units="mm" # text size is default in mm unit
)

2.1.4 ggdash more

ggdash can control any input within a programming block

myplot2 <- {
  myshape = 21 #input$shape_point
  ggplot() +
    geom_point(
      data = data4plot$point$data,
      mapping = aes(
        x = x, y = y
      ),
      fill = "#2fc1d3",
      shape = myshape,
      color = "black",
      stroke = 1,
      size = 12
    ) +
    geom_text(
      data = data4plot$text$data,
      mapping = aes(
        x = x + 
          0, #input$x_text
        y = y, label = label
      ),
      hjust = 0,
      size = 12
    )
}
myplot2

2.2 Inheritance

2.2.1 Data

Common data source can be put in ggplot():
- ALL geoms will use the same data to find aes() mapping,
- UNLESS the geom has its own data supplied.

data = data4plot$text$data

myplot <- {
  ggplot(data = data,
) +
    geom_point(
      mapping = aes(
        x = x, y = y
      ),
      fill = "#2fc1d3",
      shape = 21,
      color = "black",
      stroke = 1,
      size = 12
    ) +
    geom_text(
      # data = data4plot$text$data,
      mapping = aes(
        x = x, y = y, label = label
      ),
      hjust = 0,
      size = 12
    )
}

2.2.2 Aesthetics

myplot <- {
  ggplot(
    data = data,
    mapping = aes(x=x, y=y)
) +
    geom_point(
      fill = "#2fc1d3",
      shape = 21,
      color = "black",
      stroke = 1,
      size = 12
    ) +
    geom_text(
      mapping = aes(
        label = label
      ),
      hjust = 0,
      size = 12
    )
}
myplot

plot_hightlightCountries <- function(data) {
  myplot <- {
    ggplot(data = data,
  ) +
      geom_point(
        mapping = aes(
          x = x, y = y
        ),
        fill = "#2fc1d3",
        shape = 21,
        color = "black",
        stroke = 1,
        size = 12
      ) +
      geom_text(
        # data = data4plot$text$data,
        mapping = aes(
          x = x, y = y, label = label
        ),
        hjust = 0,
        size = 12
      )
  }
  myplot
}

2.2.3 Stop data inheritance

When a geom has its own data supplied, data inheritance stops.

nonhighlightedCountries <-
  data.frame(
    x=c(48, 52, 53),
    y=c(48, 53, 54)
  )
highlightedCountries <- data4plot$text$data

plot_allCountries <- 
{
  plot_hightlightCountries(highlightedCountries) +
    geom_point(
      data = nonhighlightedCountries,
      aes(
        x=x,
        y=y
      ),
      size = 12,
      color= "#a4dce6"
    )
}

plot_allCountries

plot_allCountries <- function(highlightedCountries, nonhighlightedCountries) {
  plot_hightlightCountries(highlightedCountries) +
    geom_point(
      data = nonhighlightedCountries,
      aes(
        x=x,
        y=y
      ),
      size = 12,
      color= "#a4dce6"
    )
}

2.2.4 Finish panel

plot_obsessionWithBorders <- {
  plot_allCountries(
    hightlightedCountries = highlightedCountries,
    unhighlightedCountries = unhighlightedCountries
  )+
  geom_abline(
    aes(
      slope=1,
      intercept=0
    ),
    color = "#549ab6",
    linetype = 2,
    size=2
  )
}
plot_obsessionWithBorders

plot_obsessionWithBorders +
  geom_abline(
    aes(
      slope=1,
      intercept=0
    ),
    color = "#549ab6",
    linetype = 2,
    size=2
  )

hightlightXYequal <- function(plot_obsessionWithBorders) {
  plot_obsessionWithBorders +
    geom_abline(
      aes(
        slope=1,
        intercept=0
      ),
      color = "#549ab6",
      linetype = 2,
      size=2
    )
}

2.3 Graph export

The easiest way to export the graph as you see on gg$dash() is to right click the image on the shiny dashboard and choose save image as to export the image.