Data visualization

# Data visualization
## Lecture 2
### Louis SIRUGUE
### M1 APE - Fall 2022

---

### Quick reminder

#### 1. Import data

```r
fb <- read.csv("C:/User/Documents/ligue1.csv", encoding = "UTF-8")
```

#### 2. Class

```r
is.numeric("1.6180339") # What would be the output?
```

```
## [1] FALSE
```

#### 3. Subsetting

```r
fb$Home[3]
```

```
## [1] "Troyes"
```

---

### Quick reminder

#### 4. Packages

```r
library(dplyr)
```

#### 5. The dplyr grammar

.left-column[
<table class="table table-hover table-condensed" style="width: auto !important; margin-left: auto; margin-right: auto;">
<caption></caption>
 <thead>
 <tr>
 <th style="text-align:left;"> Function </th>
 <th style="text-align:left;"> Meaning </th>
 </tr>
 </thead>
<tbody>
 <tr>
 <td style="text-align:left;"> mutate() </td>
 <td style="text-align:left;"> Modify or create a variable </td>
 </tr>
 <tr>
 <td style="text-align:left;"> select() </td>
 <td style="text-align:left;"> Keep a subset of variables </td>
 </tr>
 <tr>
 <td style="text-align:left;"> filter() </td>
 <td style="text-align:left;"> Keep a subset of observations </td>
 </tr>
 <tr>
 <td style="text-align:left;"> arrange() </td>
 <td style="text-align:left;"> Sort the data </td>
 </tr>
 <tr>
 <td style="text-align:left;"> group_by() </td>
 <td style="text-align:left;"> Group the data </td>
 </tr>
 <tr>
 <td style="text-align:left;"> summarise() </td>
 <td style="text-align:left;"> Summarizes variables into 1 observation per group </td>
 </tr>
</tbody>
</table>
]

---

### Warm up practice

#### 1) Import `starbucks.csv` and `View()` the data

#### 2) Inspect the structure of the data using `str()`

#### 3) Use `summarise()` to compute for each beverage category the average number of calories and the number of different declinations (there is 1 row per declination)

#### 4) Create a subset of the data called `maxcal` containing the variables `Beverage_category`, `Beverage_prep`, and `Calories`, for the 10 observations with the highest calorie values

<center>You can use the row_number() function within filter() to use the row numbers as any other variable</center>

<center><h3>You've got 10 minutes!</h3></center>

---

### Solution

#### 1) Import `starbucks.csv` and `View()` the data

```r
starbucks <- read.csv("C:/User/Documents/starbucks.csv")
View(starbucks)
```

---

### Solution

#### 1) Import `starbucks.csv` and `View()` the data

```r
starbucks <- read.csv("C:/User/Documents/starbucks.csv")
View(starbucks)
```

---

### Solution

#### 1) Import `starbucks.csv` and `View()` the data

```r
starbucks <- read.csv("C:/User/Documents/starbucks.csv")
View(starbucks)
```

<ul>
 <li>We only have one variable in which all values are separated by semicolons</li>
 <ul>
 <li>We need to set the sep argument of the function accordingly</li>
 <li>Like last time, we also need to set the encoding argument correctly</li>
 </ul>
</ul>

```r
*starbucks <- read.csv("C:/User/Documents/starbucks.csv", sep = ";", encoding = "UTF-8")
```

---

### Solution

#### 2) Inspect the structure of the data using `str()`

```r
str(starbucks)
```

```text
## 'data.frame':    242 obs. of  18 variables:
##  $ Beverage_category  : chr  "Coffee" "Coffee" "Coffee" "Coffee" ...
##  $ Beverage           : chr  "Brewed Coffee" "Brewed Coffee" "Brewed Coffee" "Brewed Coffee" ...
##  $ Beverage_prep      : chr  "Short" "Tall" "Grande" "Venti" ...
##  $ Calories           : int  3 4 5 5 70 100 70 100 150 110 ...
##  $ Total.Fat          : chr  "0.1" "0.1" "0.1" "0.1" ...
##  $ Trans.Fat          : num  0 0 0 0 0.1 2 0.4 0.2 3 0.5 ...
##  $ Saturated.Fat      : num  0 0 0 0 0 0.1 0 0 0.2 0 ...
##  $ Sodium             : int  0 0 0 0 5 15 0 5 25 0 ...
##  $ Total.Carbohydrates: int  5 10 10 10 75 85 65 120 135 105 ...
##  $ Cholesterol        : int  0 0 0 0 10 10 6 15 15 10 ...
##  $ Dietary.Fibre      : int  0 0 0 0 0 0 1 0 0 1 ...
##  $ Sugars             : int  0 0 0 0 9 9 4 14 14 6 ...
##  $ Protein            : num  0.3 0.5 1 1 6 6 5 10 10 8 ...
##  $ Vitamin.A          : chr  "0%" "0%" "0%" "0%" ...
##  $ Vitamin.C          : chr  "0%" "0%" "0%" "0%" ...
##  . ...                . ...  ...  ...  ...  ...  ...   
```

---

### Solution

#### 3) Use `summarise()` to compute for each beverage category the average number of calories and the number of different declinations (there is 1 row per declination)

```r
starbucks %>%
  group_by(Beverage_category) %>%
  summarise(Declinations = n(),
            Mean_cal = mean(Calories))
```

```
## # A tibble: 9 x 3
## Beverage_category Declinations Mean_cal
## <chr> <int> <dbl>
## 1 Classic Espresso Drinks 58 140. 
## 2 Coffee 4 4.25
## 3 Frappuccino® Blended Coffee 36 277. 
## 4 Frappuccino® Blended Crème 13 233. 
## 5 Frappuccino® Light Blended Coffee 12 162. 
## 6 Shaken Iced Beverages 18 114. 
## 7 Signature Espresso Drinks 40 250 
## 8 Smoothies 9 282. 
## 9 Tazo® Tea Drinks 52 177.
```

---

### Solution

#### 4) Create a subset of the data called `maxcal` containing the variables `Beverage_category`, `Beverage_prep`, and `Calories`, for the 10 observations with the highest calorie values

```r
maxcal <- starbucks %>%
 arrange(-Calories) %>%
 select(Beverage_category, Beverage_prep, Calories) %>%
 filter(row_number() <= 10)

maxcal
```

```
##              Beverage_category     Beverage_prep Calories
## 1    Signature Espresso Drinks           2% Milk      510
## 2    Signature Espresso Drinks           Soymilk      460
## 3  Frappuccino® Blended Coffee        Whole Milk      460
## 4    Signature Espresso Drinks Venti Nonfat Milk      450
## 5             Tazo® Tea Drinks           2% Milk      450
## 6  Frappuccino® Blended Coffee           Soymilk      430
## 7  Frappuccino® Blended Coffee Venti Nonfat Milk      420
## 8    Signature Espresso Drinks           2% Milk      400
## 9             Tazo® Tea Drinks           Soymilk      390
## 10 Frappuccino® Blended Coffee        Whole Milk      390
```

---

<h3>Today we learn how to plot data</h3>

<ul style = "margin-left:1.5cm;list-style: none">
 <li>1. The ggplot() function</li>
 <ul style = "list-style: none">
 <li>1.1. Basic structure</li>
 <li>1.2. Axes</li>
 <li>1.3. Theme</li>
 <li>1.4. Annotation</li>
 </ul>
</ul>

<ul style = "margin-left:1.5cm;list-style: none">
 <li>2. Adding dimensions</li>
 <ul style = "list-style: none">
 <li>2.1. More axes</li>
 <li>2.2. More facets</li>
 <li>2.3. More labels</li>
 </ul>
</ul>

]

<ul style = "margin-left:-1cm;list-style: none">
 <li>3. Types of geometry</li>
 <ul style = "list-style: none">
 <li>3.1. Points and lines</li>
 <li>3.2. Barplots and histograms</li>
 <li>3.3. Densities and boxplots</li>
 </ul>
</ul>

<ul style = "margin-left:-1cm;list-style: none">
 <li>4. How (not) to lie with graphics</li>
 <ul style = "list-style: none">
 <li>4.1. Cumulative representations</li>
 <li>4.2. Axis manipulations</li>
 <li>4.3. Interpolation</li>
 </ul>
</ul>

<ul style = "margin-left:-1cm;list-style: none"><li>5. Wrap up!</li></ul>
]

---

<h3>Today we learn how to plot data</h3>

---

### 1. The `ggplot()` function

#### 1.1. Basic structure

<ul>
 <li>Let's use ggplot on data from the <a href = "https://wid.world/">World Inequality database</a></li>
</ul>

```r
wid <- read.csv("C:/User/Documents/wid.csv")
```

```r
str(wid)
```

```
## 'data.frame':	1610 obs. of  6 variables:
##  $ country  : chr  "Algeria" "Algeria" "Algeria" "Algeria" ...
##  $ continent: chr  "Africa" "Africa" "Africa" "Africa" ...
##  $ year     : int  2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 ...
##  $ fshare   : num  0.0992 0.112 0.1201 0.1206 0.116 ...
##  $ top1     : num  0.1003 0.0991 0.0991 0.0991 0.0991 ...
##  $ inc_head : num  12611 12620 12634 12532 12546 ...
```

<ul>
 <li>It contains 1610 observations and 6 variables:</li>
 <ul>
 <li>continent/country/year: Observation level</li>
 <li>f_share: Female labor income share</li>
 <li>top1: Top 1% income share</li>
 <li>inc_head: Per adult national income</li>
 </ul>
</ul>

---

### 1. The `ggplot()` function

#### 1.1. Basic structure

<ul>
 <li>ggplot() from ggplot2 is what we're gonna use for all our plots</li>
 <li>It takes the following core arguments:</li>
 <ul>
 <li>Data: the values to plot</li>
 <li>Mapping (aes, for aesthetics): the structure of the plot</li>
 <li>Geometry: the type of plot</li>
 </ul>
</ul>

<ul>
 <li>Data and mapping should be specified within the parentheses</li>
 <li>Geometry and any other element should be added with a + sign</li>
</ul>

```r
ggplot(data, aes()) + geometry + anything_else
```

* You can also **apply** the `ggplot()` function to your data with a **pipe**

```r
data %>% ggplot(aes()) + geometry
```

---

### 1. The `ggplot()` function

#### 1.1. Basic structure

```r
ggplot(wid)                                         # Data
                                                    #
```

.left-column[
<img src="slides_files/figure-html/unnamed-chunk-29-1.png" width="90%" style="display: block; margin: auto;" />
]

<ul style = "margin-left:-1cm">
 <li>We assigned data to ggplot()</li>
 <ul>
 <li>But our plot is empty</li>
 </ul>
</ul>

]

---

### 1. The `ggplot()` function

#### 1.1. Basic structure

```r
ggplot(wid, aes(x = inc_head, y = top1))            # Data & aesthetics
                                                    #
```

.left-column[
<img src="slides_files/figure-html/unnamed-chunk-31-1.png" width="90%" style="display: block; margin: auto;" />
]

<ul style = "margin-left:-1cm">
 <li>We assigned data to ggplot()</li>
 <ul>
 <li>But our plot is empty</li>
 </ul>
</ul>

<ul style = "margin-left:-1cm">
 <li>We assigned variables to axes</li>
 <ul>
 <li>But still nothing</li>
 </ul>
</ul>

]
---

### 1. The `ggplot()` function

#### 1.1. Basic structure

```r
ggplot(wid, aes(x = inc_head, y = top1)) +          # Data & aesthetics
  geom_point()                                      # Geometry
```

.left-column[
<img src="slides_files/figure-html/unnamed-chunk-33-1.png" width="90%" style="display: block; margin: auto;" />
]

<ul style = "margin-left:-1cm">
 <li>We assigned data to ggplot()</li>
 <ul>
 <li>But our plot is empty</li>
 </ul>
</ul>

<ul style = "margin-left:-1cm">
 <li>We assigned variables to axes</li>
 <ul>
 <li>But still nothing</li>
 </ul>
</ul>

<ul style = "margin-left:-1cm">
 <li>We need a geometry</li>
 <ul>
 <li>Points for instance</li>
 </ul>
</ul>

]

---

### 1. The `ggplot()` function

#### 1.1. Basic structure

<ul>
 <li>You can the save the plot using the ggsave() function</li>
 <ul>
 <li>You just need to specify the output destination and it will save what is in your plot panel</li>
 </ul>
</ul>

```r
ggsave("C:/User/Documents/wid.png")
```

<ul>
 <li>You can also modify the following options, which take the parameters of your plot panel if unspecified:</li>
 <ul>
 <li>plot: ggplot object</li>
 <li>width: width of the plot</li>
 <li>height: height of the plot</li>
 <li>unit: unit of the plot size ("in", "cm", "mm", "px")</li>
 <li>dpi: pixel density, default to 300px/in</li>
 </ul>
</ul>

```r
ggsave("wid.png", plot = last_plot(), width = 16, height = 9, unit = "cm", dpi = 900)
```

---

### 1. The `ggplot()` function

#### 1.2. Axes

<ul>
 <li>Axes can be modified with scale functions, whose names depend on:</li>
 <ul>
 <li>The axis to modify</li>
 <li>The type of variable assigned to the axis</li>
 </ul>
</ul>

<table class="table table-hover table-condensed" style="width: auto !important; margin-left: auto; margin-right: auto;">
<caption>Basic scale functions</caption>
 <thead>
 <tr>
 <th style="text-align:left;text-align: center;"> Axis </th>
 <th style="text-align:left;text-align: center;"> x-axis </th>
 <th style="text-align:left;text-align: center;"> y-axis </th>
 </tr>
 </thead>
<tbody>
 <tr>
 <td style="text-align:left;"> Continuous </td>
 <td style="text-align:left;"> scale_x_continuous() </td>
 <td style="text-align:left;"> scale_y_continuous() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Discrete </td>
 <td style="text-align:left;"> scale_x_discrete() </td>
 <td style="text-align:left;"> scale_y_discrete() </td>
 </tr>
</tbody>
</table>

<ul>
 <li> The following parameters can be modified in these scale functions:</li>
 <ul>
 <li> name: The label of the corresponding axis </li>
 <li> limits: Where the axis should start and end </li>
 <li> breaks: Where to put ticks and values on the axis </li>
 </ul>
</ul>

---

### 1. The `ggplot()` function

#### 1.2. Axes

```r
ggplot(wid, aes(x = inc_head, y = top1)) + geom_point()                 # Basic structure
                                                                        #
```

---

### 1. The `ggplot()` function

#### 1.2. Axes

```r
ggplot(wid, aes(x = inc_head, y = top1)) + geom_point() +               # Basic structure
*  scale_x_continuous(name = "Income per adult", limits = c(0, 150000))  # Scale function
```

---

### 1. The `ggplot()` function

#### 1.3. Theme()

<ul>
 <li>You can use one of the default R themes to easily change the layout of your plot</li>
 <ul>
 <li> ... + theme_bw()</li>
 <li> ... + theme_minimal()</li>
 <li> ... + theme_dark()</li>
 <li>You can also tune the font size inside these functions with the base_size argument</li>
 </ul>
</ul>

---

### 1. The `ggplot()` function

#### 1.3. Theme()

<ul>
 <li>You can also custom your graph using the theme() function</li>
 <ul>
 <li>It allows to custom virtually anything</li>
 <li>Enter ?theme to see the endless list of possible arguments</li>
 <li>Obviously we won't go through all of them but here are a few</li>
 </ul>
</ul>

```r
  # Basic structure
ggplot(wid, aes(x = inc_head, y = top1)) + geom_point() + 
  # Axis
  scale_x_continuous(name = "Income per adult", limits = c(0, 150000)) +
  # Theme
  theme_minimal(base_size = 14) + 
* theme(# Color of the background and of its border
*       plot.background = element_rect(fill = "#DFE6EB", colour = "#DFE6EB"),
*       # Size of the axis lines
*       axis.line = element_line(size = rel(0.8)),
*       # Color of the grid lines
*       panel.grid = element_line(color = "gray85"))  
```

---

### 1. The `ggplot()` function

#### 1.3. Theme()

---

### 1. The `ggplot()` function

#### 1.3. Theme()

<ul>
 <li>Geometries can also be modified</li>
 <ul>
 <li>alpha: opacity from 0 to 1</li>
 <li>color: color of the geometry (for geometries that are filled such as bars, it will color the border)</li>
 <li>fill: fill color for geometries such as bars</li>
 <li>size: size of the geometry</li>
 <li>shape: change shape for geometries like points</li>
 <li>linetype: solid, dashed, dotted, etc., for line geometries</li>
 <li>...</li>
 </ul>
</ul>

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 3, 
             color = "#6794A7",
             alpha = .3,
             shape = 18)  +
  theme_minimal(base_size = 14)
```

]

.pull-right[
<img src="slides_files/figure-html/unnamed-chunk-45-1.png" width="75%" style="display: block; margin: auto;" />
]

---

### 1. The `ggplot()` function

#### 1.4. Annotation

<ul>
 <li>It is sometimes useful to annotate a graph so that certain things become more salient</li>
 <ul>
 <li>Separate two groups with a dashed line</li>
 <li>Add a few words somewhere for clarity</li>
 <li>Circle a specific group of data points</li>
 <li>Add labels to data points</li>
 </ul>
</ul>

* **Straight lines** can easily be added with their respective geometry

```r
+ geom_hline(yintercept = , linetype = )
```

```r
+ geom_vline(xintercept = , linetype = )
```

* And **punctual text annotations** can be added with `annotate()`

```r
+ annotate("text", x = , y = , label = )
```

---

### 1. The `ggplot()` function

#### 1.4. Annotation: Adding lines

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 2, alpha = .3) + 
  geom_hline(yintercept = .17)
```

<img src="slides_files/figure-html/unnamed-chunk-50-1.png" width="100%" style="display: block; margin: auto;" />
]

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 2, alpha = .3) + 
  geom_vline(xintercept = 90000, 
             linetype = "dashed")
```

<img src="slides_files/figure-html/unnamed-chunk-51-1.png" width="100%" style="display: block; margin: auto;" />
]

---

### 1. The `ggplot()` function

#### 1.4. Annotation: Adding text

```r
ggplot(wid, aes(x = inc_head, y = top1)) + geom_point(size = 2, alpha = .3) + 
 annotate("text", x = 125000, y = .28, label = "Relevant info", size = 5)
```
--
<img src="slides_files/figure-html/unnamed-chunk-53-1.png" width="60%" style="display: block; margin: auto;" />

---

### 1. The `ggplot()` function

<center>Combining everything</center>

```r
ggplot(wid, aes(x = inc_head, y = top1))
                                                                                                #
                                                                                                #
                                                                                                #
                                                                                                #
                                                                                                #
                                                                                                #
                                                                                                #
```

---

### 1. The `ggplot()` function

<center>Combining everything</center>

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 3,  color = "#6794A7", alpha = .3, shape = 18)
                                                                                                 #
                                                                                                 #
                                                                                                 #
                                                                                                 #
                                                                                                 #
                                                                                                 #
```

---

### 1. The `ggplot()` function

<center>Combining everything</center>

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 3,  color = "#6794A7", alpha = .3, shape = 18) +
  geom_vline(xintercept = 90000, linetype = "dashed", size = 1, color = "#727272")
                                                                                                 #
                                                                                                 #
                                                                                                 #
                                                                                                 #
                                                                                                 #
```

---

### 1. The `ggplot()` function

<center>Combining everything</center>

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 3,  color = "#6794A7", alpha = .3, shape = 18) +
  geom_vline(xintercept = 90000, linetype = "dashed", size = 1, color = "#727272") +
  annotate("text", x = 125000, y = .2, label = "Outliers", size = 5, color = "#505050")
                                                                                                 #
                                                                                                 #
                                                                                                 #
                                                                                                 #
```

---

### 1. The `ggplot()` function

<center>Combining everything</center>

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 3,  color = "#6794A7", alpha = .3, shape = 18) +
  geom_vline(xintercept = 90000, linetype = "dashed", size = 1, color = "#727272") +
  annotate("text", x = 125000, y = .2, label = "Outliers", size = 5, color = "#505050") +
  scale_x_continuous(name = "Income per adult", limits = c(0, 150000))
                                                                                                 #
                                                                                                 #
                                                                                                 #
```

---

### 1. The `ggplot()` function

<center>Combining everything</center>

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 3,  color = "#6794A7", alpha = .3, shape = 18) +
  geom_vline(xintercept = 90000, linetype = "dashed", size = 1, color = "#727272") +
  annotate("text", x = 125000, y = .2, label = "Outliers", size = 5, color = "#505050") +
  scale_x_continuous(name = "Income per adult", limits = c(0, 150000)) +
  scale_y_continuous(name = "Top 1% inc. share", limits = c(0, .35))
                                                                                                 #
                                                                                                 #
```

---

### 1. The `ggplot()` function

<center>Combining everything</center>

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 3,  color = "#6794A7", alpha = .3, shape = 18) +
  geom_vline(xintercept = 90000, linetype = "dashed", size = 1, color = "#727272") +
  annotate("text", x = 125000, y = .2, label = "Outliers", size = 5, color = "#505050") +
  scale_x_continuous(name = "Income per adult", limits = c(0, 150000)) +
  scale_y_continuous(name = "Top 1% inc. share", limits = c(0, .35)) +
  theme_minimal(base_size = 14)
                                                                                                 #
```

---

### 1. The `ggplot()` function

<center>Combining everything</center>

```r
ggplot(wid, aes(x = inc_head, y = top1)) + 
  geom_point(size = 3,  color = "#6794A7", alpha = .3, shape = 18) +
  geom_vline(xintercept = 90000, linetype = "dashed", size = 1, color = "#727272") +
  annotate("text", x = 125000, y = .2, label = "Outliers", size = 5, color = "#505050") +
  scale_x_continuous(name = "Income per adult", limits = c(0, 150000)) +
  scale_y_continuous(name = "Top 1% inc. share", limits = c(0, .35)) +
  theme_minimal(base_size = 14) + 
  theme(plot.background = element_rect(fill = "#DFE6EB", colour = "#DFE6EB")) 
```

---

<h3>Overview</h3>

<ul style = "margin-left:1.5cm;list-style: none">
 <li>1. The ggplot() function &#10004;</li>
 <ul style = "list-style: none">
 <li>1.1. Basic structure</li>
 <li>1.2. Axes</li>
 <li>1.3. Theme</li>
 <li>1.4. Annotation</li>
 </ul>
</ul>

]

<ul style = "margin-left:-1cm;list-style: none"><li>5. Wrap up!</li></ul>
]

---

<h3>Overview</h3>

]

---

### 2. Adding dimensions

#### 2.1. More axes

<ul>
 <li>In some cases you may want to convey information using other means than position on an axis</li>
 <ul>
 <li>The color, size, or shape of a geometry can be used to represent a third variable</li>
 </ul>
</ul>

<ul>
 <li>We can assign different colors to different points depending on the associated continent</li>
 <ul>
 <li>Continent should be assigned to the "color axis" in aes()</li>
 </ul>
</ul>

```r
ggplot(wid, aes(x = inc_head, y = top1, color = continent)) + geom_point(alpha = .3) 
```

---

### 2. Adding dimensions

#### 2.1. More axes

* If the variable assigned to the color axis is continuous, a color gradient will be used
 
--

```r
ggplot(wid, aes(x = inc_head, y = top1, color = fshare)) + geom_point(alpha = .3)
```

---

### 2. Adding dimensions

#### 2.1. More axes

<ul>
 <li>Because there is no proper "color axis", a legend is generated</li>
 <ul>
 <li>It can be seen as a "color" axis, just like the x- and y-axis</li>
 <li>And should then be modified with a scale function</li>
 </ul>
</ul>

```r
plot + scale_color_manual(
 name = "Title", values = c("red", "blue")
)
```
]
.pull-right[
<center>Continuous color variable</center>

```r
plot + scale_color_gradient(
  name = "Title", low = "red", high = "blue"
)
```
]

<ul>
 <li>But color is not the only property that can be used as a dimension, you can use:</li>
 <ul>
 <li>size, shape, alpha, ...</li>
 <li>fill, linetype, ..., for relevant geometries</li>
 </ul>
</ul>

---

.pull-left[
<img src="slides_files/figure-html/unnamed-chunk-74-1.png" width="100%" style="display: block; margin: auto;" />

<img src="slides_files/figure-html/unnamed-chunk-75-1.png" width="100%" style="display: block; margin: auto;" />
]

.pull-right[
<img src="slides_files/figure-html/unnamed-chunk-76-1.png" width="100%" style="display: block; margin: auto;" />

<img src="slides_files/figure-html/unnamed-chunk-77-1.png" width="100%" style="display: block; margin: auto;" />
]

---

### 2. Adding dimensions

#### 2.2. More facets

<ul>
 <li>Another way to distinguish groups is to divide the plot into facets</li>
 <ul>
 <li>To do so, indicate your faceting variable into the facet_wrap() function</li>
 </ul>
</ul>

* In `facet_wrap()`, the faceting variable must be preceded by a tilde as the first argument:

```r
ggplot(wid, aes(x = inc_head, y = top1)) + geom_point() +
* facet_wrap(~continent)
```

<ul>
 <li>You can then choose the facet arrangement:</li>
 <ul>
 <li>nrow to indicate the number of rows</li>
 <li>ncol to indicate the number of columns</li>
 </ul>
</ul>

.pull-left[

<ul>
 <li>As well as which scale should be:</li>
 <ul>
 <li>free: adjusted separately to each facet</li>
 <li>fixed: common to all facets</li>
 </ul>
</ul>
]

.pull-right[

<table class="table table-hover table-condensed" style="width: auto !important; margin-left: auto; margin-right: auto;">
<caption>scales argument in facet_wrap()</caption>
 <thead>
 <tr>
 <th style="text-align:left;text-align: center;"> </th>
 <th style="text-align:left;text-align: center;"> x fixed </th>
 <th style="text-align:left;text-align: center;"> x free </th>
 </tr>
 </thead>
<tbody>
 <tr>
 <td style="text-align:left;font-weight: bold;"> y fixed </td>
 <td style="text-align:left;"> scales = "fixed" </td>
 <td style="text-align:left;"> scales = "free_x" </td>
 </tr>
 <tr>
 <td style="text-align:left;font-weight: bold;"> y free </td>
 <td style="text-align:left;"> scales = "free_y" </td>
 <td style="text-align:left;"> scales = "free" </td>
 </tr>
</tbody>
</table>
]

---

### 2. Adding dimensions

#### 2.2. More facets

```r
ggplot(wid, aes(x = inc_head, y = top1)) + geom_point(alpha = .3) +
  facet_wrap(~continent, ncol = 3, scales = "free_x")
```

---

### 2. Adding dimensions

#### 2.3. More labels

<ul>
 <li>The last dimension I want to mention is the label axis</li>
 <ul>
 <li>When using geom_text() instead of geom_point(), it will plot the corresponding text instead of points</li>
 </ul>
</ul>

```r
ggplot(wid %>% filter(year == 2019 & continent == "Europe"), # subset so that we can see something
```

--

```r
  aes(x = inc_head, y = top1, label = country)) + geom_text(alpha = .6)
```

---

### Practice

#### 1) Reproduce this graph with the `starbucks` dataset

<center><h3>You've got 10 minutes!</h3></center>

---

### Solution

```r
ggplot(starbucks, 
       aes(x = Calories, y = Cholesterol)) 
                                                                                                 #
                                                                                                 #
                                                                                                 #
```

---

### Solution

```r
ggplot(starbucks, 
       aes(x = Calories, y = Cholesterol)) +
  geom_point()
                                                                                                 #
                                                                                                 #
```

---

### Solution

```r
ggplot(starbucks, 
       aes(x = Calories, y = Cholesterol, size = Trans.Fat, color = Sugars)) +
  geom_point()
                                                                                                 #
                                                                                                 #
```

---

### Solution

```r
ggplot(starbucks, 
       aes(x = Calories, y = Cholesterol, size = Trans.Fat, color = Sugars)) +
  geom_point(alpha = .3) + 
  scale_color_gradient(low = "green", high = "red")
                                                                                                 #
```

---

### Solution

```r
ggplot(starbucks, 
       aes(x = Calories, y = Cholesterol, size = Trans.Fat, color = Sugars)) +
  geom_point(alpha = .3) + 
  scale_color_gradient(low = "green", high = "red") +
  theme_minimal(base_size = 14)
```

---

<h3>Overview</h3>

<ul style = "margin-left:1.5cm;list-style: none">
 <li>2. Adding dimensions &#10004;</li>
 <ul style = "list-style: none">
 <li>2.1. More axes</li>
 <li>2.2. More facets</li>
 <li>2.3. More labels</li>
 </ul>
</ul>

]

<ul style = "margin-left:-1cm;list-style: none"><li>5. Wrap up!</li></ul>
]

---

<h3>Overview</h3>

]

---

### 3. Types of geometry

#### 3.1. Points and lines

<ul>
 <li>So far we only represented scatterplots, but many other geometries can be used</li>
 <ul>
 <li>For instance, lines are particularly suited for evolutions over time</li>
 </ul>
</ul>

```r
ggplot(wid %>% filter(country == "USA"), aes(x = year, y = top1)) + 
  geom_point() + geom_line()
```

---

### 3. Types of geometry

#### 3.2. Barplots and histograms

<ul>
 <li>Barplots however are great for categorical $x$ variables and continuous $y$ variables</li>
 <ul>
 <li>Setting the stat argument to "identity" allows to display the corresponding y value</li>
 </ul>
</ul>

```r
ggplot(wid %>% filter(continent == "South America" & year == 2019), 
       aes(x = country, y = top1)) + geom_bar(stat = "identity")
```

---

### 3. Types of geometry

#### 3.2. Barplots and histograms

<ul>
 <li>Note that you can reorder the bars according to their y value using the reorder() function</li>
</ul>

```r
ggplot(wid %>% filter(continent == "South America" & year == 2019), 
       aes(x = reorder(country, top1), y = top1)) + geom_bar(stat = "identity")
```

---

### 3. Types of geometry

#### 3.2. Barplots and histograms

<ul>
 <li>You can also set stat to "count" to plot the number of observations per category</li>
 <ul>
 <li>In that case, no variable should be assigned to the y axis</li>
 </ul>
</ul>

```r
ggplot(wid, aes(x = continent)) + geom_bar(stat = "count")
```

---

### 3. Types of geometry

#### 3.2. Barplots and histograms

<ul>
 <li>Finally, histograms can be used to describe the distribution of a continuous variable</li>
 <ul>
 <li>You can tune the bin width with binwidth or the number of bins with bins</li>
 </ul>
</ul>

```r
ggplot(wid %>% filter(year == 2019), aes(x = fshare)) + 
  geom_histogram(bins = 20, color = "white", fill = "steelblue")
```

---

### 3. Types of geometry

#### 3.3. Densities and boxplots

<ul>
 <li>The continuous equivalent of the histogram is the density</li>
 <ul>
 <li>Similarly you can tune the bandwidth with the bw argument (don't do it)</li>
 </ul>
</ul>

```r
ggplot(wid %>% filter(year == 2019), aes(x = fshare)) + 
    geom_density(color = "white", fill = "steelblue")
```

---

### 3. Types of geometry

#### 3.3. Densities and boxplots

<ul>
 <li>A handy geometry to plot densities for different groups is the violin</li>
 <ul>
 <li>Note that the grouping variable should be assigned to the $x$ axis</li>
 </ul>
</ul>

```r
ggplot(wid %>% filter(year == 2019), aes(x = continent, y = fshare)) + 
    geom_violin(color = "white", fill = "steelblue") 
```

---

### 3. Types of geometry

#### 3.3. Densities and boxplots

<ul>
 <li>Violins are particularly interesting when combined with boxplots</li>
 <ul>
 <li>When overlaying these geometries, make sure to tune the width and opacity appropriately</li>
 </ul>
</ul>

```r
ggplot(wid %>% filter(year == 2019), aes(x = continent, y = fshare)) + 
    geom_violin(fill = "steelblue", alpha = .4) + geom_boxplot(width = .1, alpha = .4)
```

---

### 3. Types of geometry

#### 3.3. Densities and boxplots

<ul>
 <li>This is how boxplots are constructed:</li>
</ul>

---

### 3. Types of geometry

#### 3.3. Densities and boxplots

<ul>
 <li>This is how boxplots are constructed:</li>
</ul>

---

### 3. Types of geometry

#### 3.3. Densities and boxplots

<ul>
 <li>This is how boxplots are constructed:</li>
</ul>

---

### 3. Types of geometry

#### 3.3. Densities and boxplots

<ul>
 <li>This is how boxplots are constructed:</li>
</ul>

---

<h3>Overview</h3>

]

<ul style = "margin-left:-1cm;list-style: none">
 <li>3. Types of geometry &#10004;</li>
 <ul style = "list-style: none">
 <li>3.1. Points and lines</li>
 <li>3.2. Barplots and histograms</li>
 <li>3.3. Densities and boxplots</li>
 </ul>
</ul>

<ul style = "margin-left:-1cm;list-style: none"><li>5. Wrap up!</li></ul>
]

---

<h3>Overview</h3>

]

---

### 4. How (not) to lie with graphics

#### 4.1. Cumulative representations

.left-column[
<center><img src = "trump.png" width = "720"/></center>
]
 
.right-column[

<center>Donald Trump during his daily coronavirus task force briefing on April 6, 2020</center>

The legend indicates: ''>1,790,000 tests completed through April 5''
]

---

### 4. How (not) to lie with graphics

#### 4.1. Cumulative representations

.left-column[
<img src="slides_files/figure-html/unnamed-chunk-116-1.png" width="100%" style="display: block; margin: auto;" />
]

.right-column[

<center>Let's take a closer look</center>
 

<center>''>1,790,000 tests completed through April 5''</center>
 

<center>Isn't there something tricky here?</center>
]
---

### 4. How (not) to lie with graphics

#### 4.1. Cumulative representations

.left-column[
<img src="slides_files/figure-html/unnamed-chunk-117-1.png" width="100%" style="display: block; margin: auto;" />
]

<ul><li>This makes it look like an exponential progression</li></ul>
 

<ul><li>While the daily number of tests actually did not increase that exponentially</li></ul>
]

---

### 4. How (not) to lie with graphics

#### 4.2. Axis manipulations

.left-column[
<img src="slides_files/figure-html/unnamed-chunk-118-1.png" width="100%" style="display: block; margin: auto;" />
]

---

### 4. How (not) to lie with graphics

#### 4.2. Axis manipulations

.left-column[
<img src="slides_files/figure-html/unnamed-chunk-119-1.png" width="100%" style="display: block; margin: auto;" />
]

.right-column[

Same data, but starting from 0
 

&#10140; Zooming or unzooming on a graph can be very misleading
]

---

### 4. How (not) to lie with graphics

#### 4.2. Axis manipulations

---

### 4. How (not) to lie with graphics

#### 4.2. Axis manipulations

.pull-left[
<img src="slides_files/figure-html/unnamed-chunk-120-1.png" width="90%" style="display: block; margin: auto;" />
<center>Misleading</center>
]

.pull-right[
<img src="slides_files/figure-html/unnamed-chunk-121-1.png" width="90%" style="display: block; margin: auto;" />
<center>Not misleading</center>
]

---

### 4. How (not) to lie with graphics

#### 4.2. Axis manipulations

* But in this case which is the most adequate representation?

.pull-left[
<img src="slides_files/figure-html/unnamed-chunk-122-1.png" width="100%" style="display: block; margin: auto;" />
]

.pull-right[
<img src="slides_files/figure-html/unnamed-chunk-123-1.png" width="100%" style="display: block; margin: auto;" />
]

---

### 4. How (not) to lie with graphics

#### 4.2. Axis manipulations

* There is no universal rule, but always pay attention to axes and scales

<img src="slides_files/figure-html/unnamed-chunk-124-1.png" width="95%" style="display: block; margin: auto;" />

<img src="slides_files/figure-html/unnamed-chunk-125-1.png" width="95%" style="display: block; margin: auto;" />
]

<img src="slides_files/figure-html/unnamed-chunk-126-1.png" width="95%" style="display: block; margin: auto;" />

<img src="slides_files/figure-html/unnamed-chunk-127-1.png" width="95%" style="display: block; margin: auto;" />
]

---

### 4. How (not) to lie with graphics

#### 4.2. Axis manipulations

<ul>
 <li>&#9888; Be very careful with double axes &#9888;</li>
 <ul>
 <li>You can make them tell basically everything</li>
 </ul>
</ul>

---

### 4. How (not) to lie with graphics

#### 4.2. Axis manipulations

<ul>
 <li>Be careful with free scales in facet_wrap() as well</li>
 <ul>
 <li>It can make things look more homogeneous than they actually are</li>
 </ul>
</ul>

---

### 4. How (not) to lie with graphics

#### 4.3. Interpolation

* Here is the **previous graph** on the tax increase using a **line geometry**
 
.pull-left[
<img src="slides_files/figure-html/unnamed-chunk-133-1.png" width="100%" style="display: block; margin: auto;" />
]

<center>But only two of them are correct</center>
]

---

### 4. How (not) to lie with graphics

#### 4.3. Interpolation

<ul>
 <li>This figure also has finitely many actual data points but feels more natural</li>
 <ul>
 <li>This is because values are sufficiently close to each other to be considered as continuous</li>
 </ul>
</ul>

---

### 4. How (not) to lie with graphics

#### 4.3. Interpolation

<ul>
 <li>There is no rule either on when lines should be used or not</li>
 <ul>
 <li>But the observation level should be clear on the graph</li>
 </ul>
</ul>

.pull-left[
<img src="slides_files/figure-html/unnamed-chunk-135-1.png" width="100%" style="display: block; margin: auto;" />

]

.pull-right[
<img src="slides_files/figure-html/unnamed-chunk-136-1.png" width="100%" style="display: block; margin: auto;" />
]

---
<h3>Overview</h3>

]

<ul style = "margin-left:-1cm;list-style: none">
 <li>4. How (not) to lie with graphics &#10004;</li>
 <ul style = "list-style: none">
 <li>4.1. Cumulative representations</li>
 <li>4.2. Axis manipulations</li>
 <li>4.3. Interpolation</li>
 </ul>
</ul>

<ul style = "margin-left:-1cm;list-style: none"><li>5. Wrap up!</li></ul>
]

---

### 5. Wrap up!

<center><h4> The 3 core components of the ggplot() function </h4></center>

<table class="table table-hover table-condensed" style="width: auto !important; margin-left: auto; margin-right: auto;">
<caption></caption>
 <thead>
 <tr>
 <th style="text-align:left;"> Component </th>
 <th style="text-align:center;"> Contribution </th>
 <th style="text-align:center;"> Implementation </th>
 </tr>
 </thead>
<tbody>
 <tr>
 <td style="text-align:left;"> Data </td>
 <td style="text-align:center;"> Underlying values </td>
 <td style="text-align:center;"> ggplot(data, | data %&gt;% ggplot(., </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Mapping </td>
 <td style="text-align:center;"> Axis assignment </td>
 <td style="text-align:center;"> aes(x = V1, y = V2, ...)) </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Geometry </td>
 <td style="text-align:center;"> Type of plot </td>
 <td style="text-align:center;"> + geom_point() + geom_line() + ... </td>
 </tr>
</tbody>
</table>

* Any **other element** should be added with a **`+` sign**

```r
ggplot(data, aes(x = V1, y = V2)) + 
  geom_point() + geom_line() +
  anything_else()
```

---

### 5. Wrap up!

.pull-left[

<center><h4> Main customization tools </h4></center>
<table class="table table-hover table-condensed" style="width: auto !important; margin-left: auto; margin-right: auto;">
<caption></caption>
 <thead>
 <tr>
 <th style="text-align:left;"> Item to customize </th>
 <th style="text-align:left;"> Main functions </th>
 </tr>
 </thead>
<tbody>
 <tr>
 <td style="text-align:left;"> Axes </td>
 <td style="text-align:left;"> scale_[x/y]_[continuous/discrete] </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Baseline theme </td>
 <td style="text-align:left;"> theme_[void/minimal/.../dark]() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Annotations </td>
 <td style="text-align:left;"> geom_[[h/v]line/text](), annotate() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Theme </td>
 <td style="text-align:left;"> theme(axis.[line/ticks].[x/y] = ..., </td>
 </tr>
</tbody>
</table>
]

.pull-right[
<center><h4> Main types of geometry </h4></center>
<table class="table table-hover table-condensed" style="width: auto !important; margin-left: auto; margin-right: auto;">
<caption></caption>
 <thead>
 <tr>
 <th style="text-align:left;"> Geometry </th>
 <th style="text-align:center;"> Function </th>
 </tr>
 </thead>
<tbody>
 <tr>
 <td style="text-align:left;"> Bar plot </td>
 <td style="text-align:center;"> geom_bar() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Histogram </td>
 <td style="text-align:center;"> geom_histogram() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Area </td>
 <td style="text-align:center;"> geom_area() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Line </td>
 <td style="text-align:center;"> geom_line() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Density </td>
 <td style="text-align:center;"> geom_density() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Boxplot </td>
 <td style="text-align:center;"> geom_boxplot() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Violin </td>
 <td style="text-align:center;"> geom_violin() </td>
 </tr>
 <tr>
 <td style="text-align:left;"> Scatter plot </td>
 <td style="text-align:center;"> geom_point() </td>
 </tr>
</tbody>
</table>
]

---

### 5. Wrap up!

<center><h4> Main types of aesthetics </h4></center>

<table class="table table-hover table-condensed" style="width: auto !important; margin-left: auto; margin-right: auto;">
<caption></caption>
 <thead>
 <tr>
 <th style="text-align:left;"> Argument </th>
 <th style="text-align:left;"> Meaning </th>
 </tr>
 </thead>
<tbody>
 <tr>
 <td style="text-align:left;"> alpha </td>
 <td style="text-align:left;"> opacity from 0 to 1 </td>
 </tr>
 <tr>
 <td style="text-align:left;"> color </td>
 <td style="text-align:left;"> color of the geometry </td>
 </tr>
 <tr>
 <td style="text-align:left;"> fill </td>
 <td style="text-align:left;"> fill color of the geometry </td>
 </tr>
 <tr>
 <td style="text-align:left;"> size </td>
 <td style="text-align:left;"> size of the geometry </td>
 </tr>
 <tr>
 <td style="text-align:left;"> shape </td>
 <td style="text-align:left;"> shape for geometries like points </td>
 </tr>
 <tr>
 <td style="text-align:left;"> linetype </td>
 <td style="text-align:left;"> solid, dashed, dotted, etc. </td>
 </tr>
</tbody>
</table>
]

<ul>
 <li>If specified in the geometry</li>
 <ul>
 <li>It will apply uniformly to all the geometry</li>
 </ul>
</ul>

<ul>
 <li>If assigned to a variable in aes</li>
 <ul>
 <li>It will vary with the variable according to a scale documented in legend</li>
 </ul>
</ul>
]

```r
ggplot(data, aes(x = V1, y = V2, size = V3)) + 
  geom_point(color = "steelblue", alpha = .6)
```