03-visualize-solution.Rmd

---
title: "Visualization"
output: github_document
---

<https://github.com/maurolepore/open-science-with-r>

## Setup

The setup chunk always runs before anything else.

* Use tidyverse package.

```{r setup}
# install.packages("tidyverse")
library(tidyverse)
```

Bored? 

* Why did I comment the the call to "install.packages()"? 
* Hint: Is it polite to force-install software in someone else's computer?



## Warm up

* Explore https://ggplot2.tidyverse.org/
* Plot what you see in Usage.

```{r}
library(ggplot2)

ggplot(mpg, aes(displ, hwy, colour = class)) +
  geom_point()
```



## Data

* Read the .csv file at <https://bit.ly/03-data-ca>

```{r}
# National Parks in California
ca <- readr::read_csv("https://bit.ly/03-data-ca")
```

Bored?

* How would you write `ca` to the file "data/03-data-ca.csv"?
* How would you re-read it from "data/03-data-ca.csv"?



Copy-paste from the code to read all other files.

```{r}
# Acadia National Park
acadia <- read_csv("https://bit.ly/03-data-acadia")
# Southeast US National Parks
se <- read_csv("https://bit.ly/03-data-se")
# 2016 Visitation for all Pacific West National Parks
visit_16 <- read_csv("https://bit.ly/03-data-visit-16")
# All Nationally designated sites in Massachusetts
mass <- read_csv("https://bit.ly/03-data-mass")
```



Preview the data with `head()`.

```{r}
head(ca)
```

Bored?

* Try also: running just `ca`, `glimpse()`, `View()`.
* Compare `as_tibble(ca)` with `as.data.frame(ca)`.




## Build

* What relationship do you expect to see between year and the number of visitors?
* Plot the data `ca`. Put `year` on x and `visitors` on y.

```{r}
# Example from https://ggplot2.tidyverse.org/
# ggplot(data = mpg, aes(x = displ, y = hwy, colour = class)) +
#   geom_point()

ggplot(data = ca, aes(x = year, y = visitors)) +
  geom_point()
```



* Map the values of the variable `park_name` to the `color` aesthetics.

```{r}
ggplot(ca, aes(year, visitors, color = park_name)) +
  geom_point()
```

Bored?

* I avoided `data =`, `x =`, and `y =` but used `colour` explicitly. Why?
* When is it a good idea to put `aes()` inside `ggplot()`?
* When is it a good idea to put `aes()` inside `geom_*()`?



## Customizing

Customize this plot:

* Add a call to `labs()` to create these labels:
  * title = "Acadia National Park Visitation".
  * y = "Visitation".
  * x = "Year".
* Add a call to `theme_bw()` to change the theme to black and white.

```{r}
ggplot(ca, aes(year, visitors, color = park_name)) +
  geom_point()
```

Bored?

* There are specific functions for x, y, and title labels. Find them.
* Find and try other themes.



## Faceting

data = se (Southeast US National Parks)

* Scatter-plot `visitors` (y) by `year` (x) for the `se` dataset.
* Add `facet_wrap(~ state)` to create a panel for each state.

```{r}
ggplot(se, aes(year, visitors)) +
  geom_point() +
  facet_wrap(~ state)
```

Bored?

* What does the `scales` argument do?
* Find another `facet_*()` function.



## Geoms

### Discrete x

data = se

You can assign to a new variable part(s) of a ggplot and reuse them later.

* Why is this a poor plot?

```{r}
p <- ggplot(se, aes(park_name, visitors))
p + 
  geom_point()
```



* Change the opacity of the points with `alpha = 0.25`.
* Here I also flip the coordinate system with `coord_flip()`
* Is this a bit better?

```{r}
p + 
  geom_point(alpha = 0.25) +
  coord_flip()
```



* Here I use `geom_jitter()` instead of `geom_point()`.
* Reorder the code so it runs correctly.

```{r}
p + coord_flip() + geom_jitter()
```



* Create a boxplot.

```{r}
p + geom_boxplot() + coord_flip()
```



### Time series

data = acadia

* Create a scatter plot
* Add lines
* Add a smoothed mean

```{r}
ggplot(acadia, aes(year, visitors)) + 
  geom_point() +
  geom_line() +
  geom_smooth()
```

Bored?

* Move the smoothed mean to become the second layer. What subtle thing happened?



### Bar charts

data = visit_16

* Create a columns bar chart where the height of the columns represent `visitors` (y) per `state` (x).

```{r}
ggplot(visit_16, aes(state, visitors)) + 
  geom_col()
```



* Map the bar fill to `park_name`.

```{r}
q <- ggplot(visit_16, aes(state, visitors, fill = park_name))
q + geom_col()
```



### Position adjustments

* Now dodge the position of the bars to compare them side by side.

```{r}
my_plot <- q + geom_col(position = "dodge")
my_plot
```



## Arranging and exporting plots

data = mass

* Export `my_plot` with `ggsave()`.

```{r}
ggsave(filename = "03-my_plot.pdf", plot = my_plot)
```

Bored?

* Try to export a pdf with RStudio's Viewer panel.



### ggplotly

* Make `my_plot` interactive with the function `ggplotly()` of the package plotly .

```{r}
if (interactive()) {
  library(plotly)
  ggplotly(my_plot)
}
```



***

# Take aways

You can use this code template to make thousands of graphs with **ggplot2**.

```{r eval = FALSE}
ggplot(data = <DATA>) +
  <GEOM_FUNCTION>(mapping = aes(<MAPPINGS>))
```