01_data_visualization.qmd

---
title: "Data visualization"
author: "Max Hachemeister"
date: last-modified
date-format: iso
format: gfm
toc: true
toc-depth: 3
---

# Prerequisites


```{r}
#| label: setup
#| include: true
#| output: false

library(tidyverse)
library(palmerpenguins)
library(ggthemes)
```

## 1.2 First steps

Variable
  : a quantity, quality, or property that you can measure

Value
  : the state of a variable when measured. May change from measurement to measurement

Observation
  : set of measurements under similar conditions. Will contain several values, each for a different variable. Sometimes also referred to as "data point"

Tabular data
 : set of values, associated with a variable and an observation. It is tidy if: Each value is in its own cell. Each variable has its own column. Each observation has its own row

> [Why now attributes, when we just now learned of variables]

Check out the `penguins` tibble:
```{r}
# Variation 1
penguins

# Variation 2
glimpse(penguins)

# Variation 3
View(penguins)

# and also run the help page
?penguins
```

## 1.2.3 Creating a ggplot

```{r}
# make an empty canvas
ggplot(data = penguins)

# map flipper_length_mm and body_mass_g to the axis
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g)
)

# add a geom to display the single observations
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g)
) + 
  geom_point()
```

### 1.2.4 Adding aesthetics and layers

> ["To achieve this, will we need to modify the aesthetic or the geom? If you guessed “in the aesthetic mapping, inside of aes()”, you’re already getting the hang of creating data visualizations with ggplot2! And if not, don’t worry."]

    > This question is somewhat misleading compared to the answer. I 'only' guessed aesthetic, and felt demotivated after reading the more specific answer.

Same plot, but this time with points colored by species:
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g, color = species)
) +
  geom_point()
```

And also add a smooth curve:
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g, color = species)
) +
  geom_point() +
  geom_smooth(method = "lm")
```

Now to have one overall line I need to write a bit different code:
```{r}
ggplot(
  data = penguins,
  mapping = aes(flipper_length_mm, y = body_mass_g)
) +
  # move the color by species down to here
  geom_point(mapping = aes(color = species)) +
  geom_smooth(method = "lm")
```

Also add shapes by species for the points, so the plot is sensible for color blindness:
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g)
) +
  geom_point(mapping = aes(color = species, shape = species)) +
  geom_smooth(method = "lm")
```

Finishing touches on the plot:
```{r}
ggplot(data = penguins,
       mapping = aes(x = flipper_length_mm, y = body_mass_g)
) +
  geom_point(mapping = aes(color = species, shape = species)) +
  geom_smooth(method = "lm") +
  labs(
    title = "Body mass and flipper length",
    subtitle = "Dimensions for Adelie, Chinstrap, and Gentoo Penguins",
    x = "Flipper length (mm)", y = "Body mass (g)",
    color = "Species", shape = "Species"
  ) + 
  scale_color_colorblind()
```

> [skipped the "mapping" for geom_point in this chunk]

### 1.2.5 Exercises

#### 1. How many rows are in `penguins` ? How many columns?
 
```{r}
glimpse(penguins)
```
 
There are 344 rows and 8 columns in `penguins`.
  
####  2. What does the `bill_depth_mm` variable in the `penguins` data frame describe?
    
```{r}
?penguins
```
    
`bill_depth_mm` describes the bill depth in millimeters of penguins

####  3. Make a scatterplot of `bill_depth_mm` vs. `bill_length_mm`. That is, make a scatterplot with `bill_depth_mm` on the y-axis and `bill_length_mm` on the x-axis. Describe the relationship between the two variables.
  
```{r}
ggplot(data = penguins,
       mapping = aes(x = bill_length_mm, y = bill_depth_mm)
) +
  geom_point()
```

The scatterplot shows a rather random pattern, with a gap between bill depth 15 to 16 mm. This indicates separate groups in the data. Let's also color the points by species and get an overall regression line to further investigate that assumption:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = bill_length_mm, y = bill_depth_mm)
) +
  geom_point(mapping = aes(color = species)) +
  geom_smooth(method = "lm") +
  # label the plot
  labs(
    title = "Bill depth and bill length",
    subtitle = "Dimensions for Adelie, Chinstrap, and Gentoo Penguins",
    x = "Bill length (mm)",
    y = "Bill depth (mm)",
    color = "Species"
  ) +
  scale_color_colorblind()
```

Now the scatterplot exhibits the different groups and their respective relationship of bill depth and bill length. However the regression line indicates a negative trend overall, but also doesn't fit the points well. Let's make a smooth line per group:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = bill_length_mm, y = bill_depth_mm,
                color = species)
) +
  geom_point() +
  geom_smooth(method = "lm") +
  labs(
    title = "Bill depth and bill length",
    subtitle = "Dimensions and regression lines for Adelie, Chinstrap and Gentoo Penguins",
    color = "Species",
    x = "Bill length (mm)",
    y = "Bill depth (mm)"
  ) +
  scale_color_colorblind()
```

The regression lines now indicate a positive relationship between bill depth and bill length within each species. Interestingly the Chinstrap Penguins seem to be separated into two groups. 

####  4. What happens if you make a scatterplot of `species` vs. `bill_depth_mm`? What might be a better choice of geom?

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = species, y = bill_depth_mm)
) +
  geom_point()
```
  
As `species` is a categorical variable, all the points are on a line for each species. So we can somewhat see that Gentoo Penguins have generally the shortest bills, there is not much more information in this plot. Which other geom would be sensible though?
    So far we the only other geom we learned is `geom_smooth`. Let's try that:
    
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = species, y = bill_depth_mm)
) +
  geom_smooth(method = "lm")
```
    
Okay that's even less informative. 
  
So let's see what other geoms come up when I type `geom_` into the console:  
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = species, y = bill_depth_mm)
) +
geom_bin_2d()
```

Okay `geom_bin2d()` seems more informative as you can see how many observations there are for each value of `bill_depth_mm`. 

I mean what we also could do, is plotting `bill_depth_mm` vs. `body_mass` and color by species, aswell as adding a regression line, like in the first example plot. Let's try that:
    
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = bill_depth_mm, y = body_mass_g)
) +
  geom_point(mapping = aes(color = species, shape = species)) +
  geom_smooth() +
  scale_color_colorblind() +
  labs(
    title = "Bill depth and body mass",
    subtitle = "Dimensions for Adelie, Chinstrap and Gentoo Penguins",
    x = "Bill depth (mm)",
    y = "Body mass (mm)"
  )
```
  Well now we see that Gentoo Penguins, while being the heaviest, seem to have generally thinner bills. The regression line is wonky as there seem to be two groups and those are not linearly related. But yeah. That should suffice for now.
    
#### 5. Why does the following give an error and how would you fix it?
  
```{r}
#| error: true

ggplot(data = penguins) +
  geom_point()
```
  
 `geom_point` expects the `x` and `y` aesthetics to be mapped. So I would map for example `bill_depth_mm` and `body_mass_g` to `x` and `y` respectively.
    
####  6. What does the `na.rm` argument do in `geom_point()`? What is the default value of the argument? Create a scatterplot where you successfully use this argument set to `TRUE`.

First check the documentation for geom_point()  
```{r}
?geom_point()
```
  
The documentation states that this argument's default value is `FALSE`. And further down it explains that this argument removes missing values either giving out a warning or not. Let's do a plot and set `na.rm` to `TRUE`:
    
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = bill_depth_mm, y = body_mass_g, color = species)
) +
  geom_point(na.rm = TRUE)
```

Ah yeah, there is no more warning in the output.
    
####  7. Add the following caption to the plot you made in the previous exercise:"Data come from the palmerpenguins package." Hint: Take a look at the documentation for `labs()`.

Check the documentation for labs  
```{r}
?labs()
```
  
According to the documentation there is an `caption` argument, so I will try this for the plot:

    
Let's make it with shapes and for colorblind aswell
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = bill_depth_mm, y = body_mass_g, 
                color = species, shape = species)
) +
  # and keep the na.rm for tidy code
  geom_point(na.rm = TRUE) +
  scale_color_colorblind() +
  labs(
    title = "Bill depth and Body mass",
    subtitle = "Dimensions for Adelie, Chinstrap and Gentoo Penguins",
    # here comes the caption
    caption = "Data comes from the palmerpenguins package.",
    x = "Bill depth (mm)",
    y = "Body mass (g)"
  )
```
    
####  8. Recreate the following visualization. What aesthetic should `bill_depth_mm` be mapped to? And should it be mapped at the global level or at the geom level?

![Example Plot](images/r4ds_1_exercise_referenceplot.png)
We see a scatterplot of `body_mass_g` and `flipper_length_mm` with an added `geom_smooth()`. Everything else is default.
Let's recreate it:
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g)
) +
  geom_point() +
  geom_smooth(method = "lm")
```

Ah now I see that there is also `bill_depth_mm` mapped to the `color` aesthetic.
Let's add that aswell:
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g,
                color = bill_depth_mm)
) +
  geom_point() +
  geom_smooth(method = "lm")
```

Okay, that seems right. There are many warnings, but the plot got rendered.

#### 9. Run this code in your head and predict what the output will look like. Then, run the code in R and check your predictions.

```{r}
#| eval: false

ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g, color = island)
) +
  geom_point() +
  geom_smooth(se = FALSE)
```

`geom_point()` and `geom_smooth()` tell me that this will be a scatterplot of `flipper_length_mm` and `body_mass_g` with a regression line. Ah and `color = island` tells me that the points will be colored by `island`, and I think that also makes a regression line for each `island` group.
Let's see:
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g, color = island)
) +
  geom_point() +
  geom_smooth(se = FALSE)
```

Ah and the `se = false` argument for `geom_smooth()` hides the gray areas around the regression lines.

#### 10. Will these two graphs look different? Why/why not?

```{r}
#| eval: false

ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g)
) +
  geom_point() +
  geom_smooth()

ggplot() +
  geom_point(
    data = penguins,
    mapping = aes(x = flipper_length_mm, y = body_mass_g)
  ) +
  geom_smooth(
    data = penguins,
    mapping = aes(x = flipper_length_mm, y = body_mass_g)
  )
```

Actually, I think both codes will get the same graph. The layers `geom_point()` and `geom_smooth()` are the same in each and from `?geom_point` I could read that geoms also takes the `data` and `mapping` arguments directly. So while the second code is more to type both are the same i guess.
Let's see:
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g)
) +
  geom_point() +
  geom_smooth()

ggplot() +
  geom_point(
    data = penguins,
    mapping = aes(x = flipper_length_mm, y = body_mass_g)
  ) +
  geom_smooth(
    data = penguins,
    mapping = aes(x = flipper_length_mm, y = body_mass_g)
  )
```

## 1.4 Visualizing distributions

### 1.4.1 Categorical Variables

*Categorical* Variables
: aka *qualitative* variables
: can take only a small set of values
: mathematical calculations don't yield practical results

>[Why not also *qualitative* variable here, like with numerical variables?]

For example `species`.
`geom_bar()` is good for that. It counts the number of observations for each 'level' of a category and sizes the bars accordingly.
```{r}
ggplot(
  data = penguins,
  mapping = aes(x = species)
) +
  geom_bar()
```

Use `fct_infreq()` to get the bars ordered by count.

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = fct_infreq(species))
) +
  geom_bar()
```

### 1.4.2 Numerical variables

*numerical variables*
: aka *quantitative* variables
: can take a wide range of numerical values
: you get sensible results when calculating with the values

Example `body_mass_g`.
`geom_histogram()` and `geom_density()` are good for that.

While the histogram is more true to the original distribution of the data, the density curve usually is more easy to get an intuition of the distribution. However the line in the plot is only a mathematical approximation, so some information is not visible.
Let's check it out:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = body_mass_g)
) +
  geom_density()
```

### 1.4.3 Exercises

#### 1. Make a barplot of `species` of `penguins`, where you assing `species` to the `y` aesthetic. How is this plot different?
>> why is scatterplot written as one and bar plot written as two words?

```{r}
ggplot(
  data = penguins,
  mapping = aes(y = species)
) +
  geom_bar()
```
Okay and the same plot with `species` on the `x` axis:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = species)
) +
  geom_bar()
```

Gotcha, so the bars either go in `x` direction, that is vertical; or they go in `y` direction, that is, horizontal.

#### 2. How are the following two plots different? Which aesthetic, `color` or `fill`, is more useful for changing the color of the bars?

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = species)
) +
  geom_bar(color = "red")

ggplot(
  data = penguins,
  mapping = aes(x = species)
) +
  geom_bar(fill = "red")
```

So `color` applies to the outline and `fill` to the area of the geom.

#### 3. What does the `bins` argument in `geom_histogram()` do?

I can remember this.
The `bins`... hold up!

So `bins` and `binwidth` are separate arguments.
`binwidth` was explained before. Its value defines how wide each bin is, in the units of the variable that is mapped to the geom.

I think with  `bins` you define how many bins in total you want the data to be sorted into.
Let's try it by making a histogram with ten bins:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = body_mass_g)
) +
  geom_histogram(bins = 10)
```

Yep, that's ten.

#### 4. Make a histogram of the `carat` variable in the `diamonds` dataset that is available when you load the tidyverse package. Experiment with different binwidths. What binwidth reveals the most interesting patterns?

Let's first take a look at the dataset, to get a sense for the scale of the `carat` and also check the documentation:
```{r}
glimpse(diamonds)

?diamonds
```


Good, it says `carat` has values from 0.2 to 5.01.
So I start with a `binwidth` of 0.1 and 0.01, to see what histograms I get.

```{r}
ggplot(
  data = diamonds,
  mapping = aes(x = carat)
) +
  geom_histogram(binwidth = 0.1)

ggplot(
  data = diamonds,
  mapping = aes(x = carat)
) +
  geom_histogram(binwidth = 0.01)
```

Interesting!
There seems to be a pattern. A lot of diamonds are around values that are a quarter of a carat.

Let's do two things.
One histogram with `binwidth` 0.25 to check if I read the pattern rigth and one with 0.005 to see whether it repeats on even lower scale.

```{r}
ggplot(
  data = diamonds,
  mapping = aes(x = carat)
) +
  geom_histogram(binwidth = 0.25)

ggplot(
  data = diamonds,
  mapping = aes(x = carat)
) +
  geom_histogram(binwidth = 0.005)
```

Ahh! 
The 0.25 did not bring something new. I misinterpreted that.
But check out the 0.005 histogram! There are some empty bins. I guess we found the minimal carat they could measure, which is 0.01.
And it seems that there are more diamonds on whole carat and quarter fractions.
Is probably a size that buyers like to see.

## 1.5 Visualizing relationships

### 1.5.1 A numerical and a categorical variable

Example, boxplot of body mass and species:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = species, y = body_mass_g)
) +
  geom_boxplot()
```

Or alternatively density plots by species:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = body_mass_g, color = species)
) +
  geom_density(linewidth = 0.75)
```

To make it more appealing, add `fill`, but make the overlap visible with the `alpha` argument:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = body_mass_g, color = species, fill = species)
) +
  geom_density(alpha = 0.5, linewidth = 0.75)
```

mapping to an aesthetic
: have the visual attribute vary based on the values of the variable

setting an aesthetic
: set the value 'manually'

### 1.5.2 Two categorical variables

Example, stacked bar plots of `island` and `species`.

First one so you can see the raw counts:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = island, fill = species)
) +
  geom_bar()
```

Second one to see the proportions, with `position = "fill"` argument:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = island, fill = species)
) +
  geom_bar(position = "fill")
```

### 1.5.3 Two numerical variables

Example, scatterplot of flipper length and body mass.
Just like in section 1.1 and 1.2

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g)
) +
  geom_point()
```

### 1.5.4 Three or more variables

Example, scatterplot of flipper length and body mass, with colors by species and shape by island.

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g,
                color = species, shape = island)
) +
  geom_point()
```

But yeah, that's pretty busy and relatively less informative.

At this point, splitting the plot into facets with `facet_wrap()` would be one solution.
Note the `~` in the facet wrap function. This is the *tilde* sign, which is used in R for *formulas*. More on that later though.

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = flipper_length_mm, y = body_mass_g, 
                color = species, shape = species)
) +
  geom_point() +
  facet_wrap(~island)
  
```

### 1.5.5 Exercises

#### 1. The `mpg` data frame that is bundled with the ggplot2 package contains 234 observations collected by the US Environmental Protection Agency on 38 car models.

  - Which variables in `mpg` are categorical?
  - Which variables are numerical?
    Hint: Type `?mpg` to read the documentation for the dataset.
  - How can you see this information when you run `mpg`?

Read the documentation, `glimpse()` the dataframe and run `mpg` directly:

```{r}
?mpg

glimpse(mpg)

mpg
```

The documentation explains the columns further, but does not list the variable type, while `glimpse()` gives a short overview of the dataframe and shows the variable type in its second row. You can find this information also when calling `mpg` directly, then it's right under each column header. You have to scroll through the columns though.

In total there are six categorical variables (basically all `<chr>` columns) and five numerical variables.

The categorical variables are `manufacturer`, `model`, `trans`, `drv`, `fl` and `class`.

The numerical variables are `displ`, `year`, `cyl`, `cty` and `hwy`.

Well come to think of it, `cyl`, which is the number of cylinders might count as a categorical variable, as there are usually a few possible values for that and arithmetic doesn't seem sensible in that case.

#### 2. Make a scatterplot of `hwy`vs. `displ` using the `mpg` data frame.
> Next, map a third numerical variable to `color`, then `size`, then both `color` and `size`, then `shape`. How do these aesthetics behave differently for categorical vs. numerical variables?

Scatterplot of `hwy` vs. `displ`.
The vs. implicates for me that you read it like 'y vs. x'. Well that's at least a convention I follow.

```{r}
ggplot(
  data = mpg,
  mapping = aes(x = displ, y = hwy)
) +
  geom_point()
```

Adding `trans`, a categorical to `color`:

```{r}
ggplot(
  data = mpg,
  mapping = aes(x = displ, y = hwy, color = trans)
) +
  geom_point()
```

Looks very busy to me. But I can make out that auto(l4) and manual(m5) seem to be the most common transmission types.

Add `cty` as a numerical variable to `size`.

```{r}
ggplot(
  data = mpg,
  mapping = aes(x = displ, y = hwy,
                color = trans, size = cty)
) +
  geom_point()
```

So it seems like it put the numerical variable into bins and then draws the point size accordingly.

Now map `cty` to both `color` and `size`:

```{r}
ggplot(
  data = mpg,
  mapping = aes(x = displ, y = hwy,
                color = cty, size = cty) 
) +
  geom_point()
```

Now we see that `color` and `size` of the points are continuous and the bins are just in the legend to help you orient.

Now map `cty` only to shape:

```{r}
#| error: true

ggplot(
  data = mpg,
  mapping = aes(x = displ, y = hwy,
                shape = cty)
) +
  geom_point()
```

Ah okay, continuous variables are not by default mapable to `shape`.

#### 3. In the scatterplot of `hwy` vs. `dspl`, what happens if you map a third variable to `linewidth`?

I will map `cty` to `linewidth`.
Interesting what will happen, as it's just points. So what does `linewidth` even apply to?
Let's see:

```{r}
#| warning: true


ggplot(
  data = mpg,
  mapping = aes(x = displ, y = hwy)
) +
  geom_point(
    mapping = aes(linewidth = cty)
  )
```

Ah, so `linewidth` is unknown for the scatterplot and therefore ignored, but at least a warning is given.

#### 4. What happens if you map the same variable to multiple aesthetics?

Let's map `hwy` to `x`, `y` and `color`:

```{r}
ggplot(
  data = mpg,
  mapping = aes(x = hwy, y = hwy, color = hwy)
) +
  geom_point()
```

Seems like it did what I asked for. It's just not much new information. We can see the distribution a bit.

#### 5. Make a scatterplot of `bill_depth_mm` vs. `bill_length_mm` and color the points by `species`.
> What does adding coloring by species reveal about the relationship between these two variables? What about faceting by `species`?

> [!note] Why is the second `species` not formated like the other occurences?

So there are two visualizations to be done. Let's start with the scatterplot:

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = bill_length_mm, y = bill_depth_mm, color = species)
) +
  geom_point()
```

The colors in the plot make out the distinction between the species. You can see that the points for each form separate groups in the value ranges of bill length and bill depth.

Let's do the facets now!
I will keep the coloring, as I think it won't harm the appeal of the plot.

```{r}
ggplot(
  data = penguins,
  mapping = aes(x = bill_length_mm, y = bill_depth_mm, color = species)
) +
  geom_point() +
  facet_wrap(~ species)
```

Well you can see the same. The species make up clusters. So Gentoo penguins have the longest but thinnest bills, while Adelie and Chinstrap penguins have similarly wide bills, even though Adelies have generally the shortest bills of all three species.

I think though that you can see the relation of those groups better when they are all in one plot, whereas the faceting makes the individual ranges more clear.

#### 6.

> Why does the following yield two separate legends? How would you fix it to combine the two legends?

Sample code:

```{r}
ggplot(
  data = penguins,
  mapping = aes(
    x = bill_length_mm, y = bill_depth_mm,
    color = species, shape = species
  )
) +
  geom_point() +
  labs(color = "Species")
```

The `labs` makes it so that `shape` and `color` get a different label for their legends. Maybe if I leave `labs` out, both will become one legend.

Let's see:

```{r}
ggplot(
  data = penguins,
  mapping = aes(
    x = bill_length_mm, y = bill_depth_mm,
    color = species, shape = species
  )
) +
  geom_point()
```

Good, that solved it.
And if I wanted to have "Species" with the capital "S" I would have to set this for both aesthetics in the `labs` function.

```{r}
ggplot(
  data = penguins,
  mapping = aes(
    x = bill_length_mm, y = bill_depth_mm,
    color = species, shape = species
  )
) +
  geom_point() +
  labs(
    color = "Species",
    shape = "Species"
  )
```

#### 7.

> Create the two following stacked bar plots. Which question can you answer with the first one? Which question can you answer with the second one?

```{r}
ggplot(penguins, aes(x = island, fill = species)) +
  geom_bar(position = "fill")

ggplot(penguins, aes(x = species, fill = island)) +
  geom_bar(position = "fill")
```

The first stacked bar plot answers the question: "What is the proportion of species among all the penguins of each island?"

The second stacked bar plot answers the question: "How are the penguins of each species distributed among the islands?"

You can roughly read it like: "What is the proportion of `fill` among all the 'observational units' for each `x`?"

So in case of the second plot it would read: "What is the proportion of `islands` among all the 'penguins' for each `species`?

Yeah, you'd have to tweak it a bit case by case.

## 1.6 Saving your plots

### 1.6.1 Exercises

#### 1.

> Run the following lines of code. Which of the two plots is saved as `mpg-plot.png`? Why?

Example code:

```{r}
ggplot(mpg, aes(x = class)) +
  geom_bar()
ggplot(mpg, aes(x = cty, y = hwy)) +
  geom_point()
ggsave("mpg-plot.png")
```

Can I display this plot somehow from the file?
Well the internet search shows some extra packages that can do that, but not natively within R-Studio. So you will have to believe me that 

I've checked the `mpg-plot.png` and it shows the scatterplot.

This reflects the mentioned behavior of `ggsave()`, saving the most recently rendered plot, which in the code above is the scatterplot, if you were to run the chunk as a whole.

#### 2.

> What do you need to change in the code above to save the plot as a PDF instead of a PNG? How could you find out what types of image files would work in `ggsave()`?

To see what `ggsave()` can do, I would check the documentation like so:

```{r}
?ggsave()
```

There it says for the argument "device", that it can do a bunch of formats, e.g. "eps", "ps", *"pdf"*, "png" and such. In the end it's also described that it guesses the format based on the `filename` extension, if not stated otherwise in the "device" argument.

So to save the plot as "pdf" I can just name the output accordingly:

```{r}
ggplot(mpg, aes(x = cty, y = hwy)) +
  geom_point()
ggsave("mpg-plot.pdf")
```

## 1.7 Common problems

Writing code that works on first try takes practice and consideration.
Mostly something won't work.
Be patient and read your code line by line.
Check the documentation. In the beginning especially the code examples for the functions.
Also the web search, or LLM can be of good help, if you're stuck.

## 1.8 Summary

I've made my first scatterplot, that would be ready to publish. With labels, shapes and colorblind coloring.

I learned the basics of `ggplot()`. How it makes up plots of different layers and how to control those elements, by either "setting" them directly, or "mapping" them to express values of a variable.

I learned several plots to visualize different relationships, for example, two numeric variables, one categorical and one numeric and such.

So i also learned about the distinction between those variable types.

In the end I learned to save a plot and how to troubleshoot my code a bit.