---
title: "Graphics and maps with R"
subtitle: "<br/>Under Hadley Wickham's reign"
author: "Marie-Pierre Etienne"
date: '2020/09/11 (updated: `r Sys.Date()`)'
institute: https://github.com/MarieEtienne/stateoftheRinRennes/
csl: ../resources/apa-no-doi-no-issue.csl
output:
  xaringan::moon_reader:
    css: [  'metropolis', 'hygge', 'mpe_pres.css']
    lib_dir: libs
    nature:
      highlightStyle: github
      highlightLines: true
      countIncrementalSlides: false
      beforeInit: '../resources/collapseoutput.js'
    includes:
      after_body: '../resources/insert-logo.html'
bibliography: visu.bib
fontsize: 10pt
params:
  child_path: ''
  setup_path: ../resources/
---

```{r setup,   cahe = FALSE, eval = TRUE, echo = FALSE, message = FALSE}
library(flipbookr)
library(RefManageR)
library(kableExtra)
source(paste0(params$setup_path, "knitr_setup.R"))
with_sol <- FALSE ## in order to control the output
with_course <- TRUE
```

```{r reference,  include=FALSE, cache=FALSE, eval = TRUE}
BibOptions(check.entries = FALSE,
           bib.style = "authoryear",
           cite.style = "alphabetic",
           style = "markdown",
           hyperlink = FALSE,
           dashed = FALSE)
myBib <- ReadBib("./visu.bib", check = FALSE)
```


# Introduction


- [R for Data Science](https://r4ds.had.co.nz/) `r Citep(myBib, "wickham2017R")`
- [R Graphics Cookbook](http://www.cookbook-r.com/)
- [ggplot2: Elegant Graphics for Data Analysis](https://ggplot2-book.org/index.html) `r Citep(myBib, "wickham2016ggplot2")` 
- [Geocomputation with R](https://geocompr.robinlovelace.net/) `r Citep(myBib, "lovelace2019geocomputation")`
- [Tutoriel de Raphaelle Momal et Marie Perrot Dockes](https://stateofther.github.io/post/fancy-plotting/)
- [Drawing beautiful maps programmatically with R](https://www.r-spatial.org/r/2018/10/25/ggplot2-sf.html)



```{r graph_bib, out.height = "150px", eval = TRUE}
book1 <- ggdraw() + draw_image("https://d33wubrfki0l68.cloudfront.net/b88ef926a004b0fce72b2526b0b5c4413666a4cb/24a30/cover.png")
book2 <-  ggdraw() + draw_image("http://www.cookbook-r.com/r_graphics_cookbook.png")
book3 <-  ggdraw() + draw_image("https://geocompr.robinlovelace.net/images/cover.png")
plot_grid(book1, book2, book3, nrow = 1)
```

---

# ggplot presentation

gg stands for Grammar of Graphics  and  ggplot2  is *based on the Grammar of Graphics `r Citep(myBib, "wilkinson2005gramma")`, that allows you to compose graphs by combining independent components* `r Citep(myBib, "wickham2016ggplot2")`.

*[The] grammar tells us that a graphic maps the*
- *data* 
- *to the aesthetic attributes (colour, shape, size)* 
- *of geometric objects (points, lines, bars)*. 
- *The plot may also include statistical transformations of the data and information about the plot’s coordinate system*.
- *Facetting can be used to plot for different subsets of the data.* 
*The combination of these independent components are what make up a graphic.*  `r Citep(myBib, "ickham2016ggplot2")`.


For coherence, we will make use of the `tidyverse` suite. Be sure to understand the `%>%` operator.


```{r load_tidyverse, message = FALSE, eval = TRUE, echo = TRUE}
library(tidyverse)
```


```{r tidy_ex}
set.seed(1)
n <- 8
data.frame(x = sample(1:10, size = n, replace = TRUE), y = 3*rnorm(n)) %>% 
  mutate(xplusy = x + y ) %>% 
  filter( xplusy > 4) %>% 
  select(x, y)
```

---

`r chunk_reveal("tidy_ex", break_type = "auto")`

---


# Presentation based on 


- [ggplot2: Elegant Graphics for Data Analysis](https://ggplot2-book.org/index.html) `r Citep(myBib, "wickham2016ggplot2")` 
- [Geocomputation with R](https://geocompr.robinlovelace.net/) `r Citep(myBib, "lovelace2019geocomputation")`
- [Tutoriel de Raphaelle Momal et Marie Perrot Dockes](https://stateofther.github.io/post/fancy-plotting/)
- [Drawing beautiful maps programmatically with R](https://www.r-spatial.org/r/2018/10/25/ggplot2-sf.html)
- [Don’t cha wish your ggplot had colours like me?](https://www.dataembassy.co.nz/Liza-colours-in-R#22)

---

# First running example:  The Palmer Penguins dataset



```{r palmer_data, eval = TRUE, echo = TRUE}
#remotes::install_github("allisonhorst/palmerpenguins")
data(penguins,package = "palmerpenguins")
penguins <- penguins %>%
  rename(bill_l = bill_length_mm, bill_d = bill_depth_mm, flip_l = flipper_length_mm, bm = body_mass_g)
penguins %>%
  print(n=2)
```


---

# The Palmer Penguins dataset : overview


```{r show_data_palmer, eval = TRUE, echo = TRUE}

penguins %>%
  group_by(species, sex, year, island) %>%
  mutate(n = n()) %>%
  summarise_if(is.numeric, mean, na.rm = TRUE) %>%
  print(n=10)
```


---
 
# Simple plots 

---
name: scatter_slide
# A scatter plot

```{r simple_scatter}
penguins %>% 
  ggplot() + 
  aes( x= bill_l, y=bill_d) + 
  geom_point()
```

---

`r chunk_reveal("simple_scatter", break_type = "auto")`

---
template: scatter_slide
count: false

## with colors 

```{r scatter_plot_species}
penguins %>% 
  ggplot() +
  aes( x= bill_l, y=bill_d) +
  geom_point() + #BREAK
  aes(col = species)  #BREAK
```

---

`r chunk_reveal("scatter_plot_species", break_type = "user")`

---
template: scatter_slide
count: false


```{r gg1_save, eval = TRUE, echo = TRUE}
gg <- penguins %>% 
  ggplot() +
  aes( x= bill_l, y=bill_d) +
  geom_point() + 
  aes(col = species)  
```


## with a color blind compliant palette


---

```{r scatter_viridis}
gg + 
  scale_color_viridis_d()
```

`r chunk_reveal("scatter_viridis", break_type = "auto")`

---
template: scatter_slide
count: false

##  with a color tribute to [Wes Anderson](https://en.wikipedia.org/wiki/Wes_Anderson) 

Following  a Tumblr blog  [Wes Anderson Palettes](https://wesandersonpalettes.tumblr.com/), Karthik Ram proposes the [wesanderson palette](https://github.com/karthik/wesanderson) on Github.


```{r scatter_define_palette_wesanderson,eval = TRUE, echo = TRUE}
## remotes::install_github("wesanderson")
color_darj <- wesanderson::wes_palette(name = "Darjeeling1")
```

---

```{r scatter_wesanderson}
gg + 
  scale_color_manual(values = color_darj)
```


`r chunk_reveal("scatter_wesanderson", break_type = "auto")`

---
name: scatter_anderson

# A scatter plot

##  with a color tribute to [Wes Anderson](https://en.wikipedia.org/wiki/Wes_Anderson) 

---
template: scatter_anderson
count: true

### Exercise

Propose a fancy graph to visualize the relation between flipper and body mass. Does this relation change over year ?


--

```{r exo_relation_cor, eval =  TRUE, echo = with_sol}
penguins %>% ggplot() +
  aes(x= bm, y = flip_l, col = as_factor(year) ) + 
  geom_point() + 
  scale_color_manual(values = wesanderson::wes_palette(name = "Darjeeling1"))

```


---
template: scatter_slide

## Some nice color palette
- [Inspired by Ghibli](https://github.com/ewenme/ghibli)
- [The yarr pirate palette](https://cran.r-project.org/web/packages/yarrr/vignettes/piratepal.html)
- [Harry Potter inspiration](https://github.com/aljrico/harrypotter)![](https://github.com/aljrico/harrypotter/blob/master/man/figures/logo.png)


---
template: scatter_anderson
count: true

## with nice labels

```{r scatter_labels_gg, eval = TRUE, echo = TRUE}
gg <- gg + scale_color_manual(values = color_darj) 
```


---

```{r scatter_labels}
gg  +
  labs( x = 'Bill length in mm') +
  labs(y = 'Bill depth in mm') +
  labs(color = "Species")
```

`r chunk_reveal("scatter_labels", break_type = "auto")`




---
name: scatter_label

# A scatter plot

##  with a color tribute to [Wes Anderson](https://en.wikipedia.org/wiki/Wes_Anderson),  nice labels



---
template: scatter_label
count: true

## and a pre-set general theme

---

```{r scatter_themelight}
gg  +
  labs( x = 'Bill length in mm') +
  labs(y = 'Bill depth in mm') +
  labs(color = "Species") + #BREAK
  theme_light() #BREAK
```


`r chunk_reveal("scatter_themelight", break_type = "user")`

---
template: scatter_label
count: false

## and another pre set general theme


---

```{r scatter_thememinimal}
gg + scale_color_manual(values = color_darj) +
  labs( x = 'Bill length in mm',  
        y = 'Bill depth in mm',  color = "Species") +
  theme_light() + #BREAK 
  theme_minimal() #BREAK 
```


`r chunk_reveal("scatter_thememinimal", break_type = "user")`

---
template: scatter_label
count: false

## and a pre set general theme


```{r scatter_set_theme, eval = TRUE, echo = TRUE}
gg <- gg + scale_color_manual(values = color_darj) +
  labs( x = 'Bill length in mm',  y = 'Bill depth in mm',  color = "Species") +
  theme_light()
```

---
template: scatter_label
count: true

## and a custom  theme

The default theme might not be the best option



```{r preset_theme, eval = TRUE, echo = FALSE}
gg
```

---

```{r scatter_below}
gg + 
  theme(legend.position="bottom")
```

`r chunk_reveal("scatter_below", break_type = "auto")`

---
template: scatter_label
count: false

## and a custom  theme

The legend position can also be specified by a vector ` c(x,y)`.  Their values should be between 0 and 1. c(0,0) corresponds to the “bottom left” and c(1,1) corresponds to the “top right” position. We may want to change the size.


```{r scatter_within, eval = TRUE, echo = TRUE}
gg + 
  theme(legend.position=c(.9, .6))
```


---
template: scatter_label
count: false

## and a custom  theme

```{r scatter_custom, eval = TRUE, echo = TRUE}
gg + theme(legend.position=c(.9, .6),
           text = element_text(size = 10, face = "italic"),
           axis.text.x = element_text(angle=90, hjust=1),
           legend.text = element_text(size = 9, face = 'plain'),
           legend.title = element_text(size = 11, face = "bold") )

```

---
template: scatter_label
count: true

## and fitted statistical models

```{r}
gg
```


---


```{r stat_info_lm}
gg + 
  geom_smooth(method = 'lm', se = FALSE) +
  geom_smooth(method = 'loess', se = TRUE)
```




`r chunk_reveal("stat_info_lm", break_type = "auto")`


---
template: scatter_anderson
count: true

### Exercise

Propose a fancy graph to visualize the relation between flipper and body mass. Does this relation change over year ?


--

```{r exo_relation_cor2, eval =  TRUE, echo = with_sol}
penguins %>% ggplot() + 
  aes(x= bm, y = flip_l, col = as_factor(year) ) + 
  geom_point() + geom_smooth(method = 'lm') +
  scale_color_manual(values = wesanderson::wes_palette(name = "Darjeeling1"))

```

---

# A scatter plot

##  with a color tribute to [Wes Anderson](https://en.wikipedia.org/wiki/Wes_Anderson),  nice labels, nice custom theme

To keep the theme preferences for all plots to come

```{r theme_setup}

penguins %>% 
  ggplot() +
  aes( x= bill_l, y=bill_d, col = species) +
  geom_point()  #BREAK

```

---

`r chunk_reveal("theme_setup", break_type = "user")`



```{r theme_setup_2}

theme_set(theme_light())
theme_update(legend.position="bottom") #BREAK
penguins %>% 
  ggplot() +
  aes( x= bill_l, y=bill_d, col = species) +
  geom_point() #BREAK
```


---

`r chunk_reveal("theme_setup_2", break_type = "user")`


---
# Many other type of plots 


```{r simple_hist}
penguins %>%
  ggplot() + aes(x = bill_l) + geom_histogram(position = 'identity') +
  labs( x = 'Bill length in mm') +
  theme_minimal()
```


---

# Many other types of plots 

## Histogram - exercise

Add the color according to species and use the `Zissou1` palette from `wesanderson` package.

```{r color_hist,   echo = TRUE}
 penguins %>%  ...
```


---


```{r color_hist_cor}
penguins %>%
  ggplot() + aes(x = bill_l, fill = species) + geom_histogram() +
  labs( x = 'Bill length in mm') +
  scale_fill_manual(values = wesanderson::wes_palette('Zissou1', n = 3)) 
```


`r chunk_reveal("color_hist_cor", break_type = "auto")`

---


# Many other types of plots 

## Comparing plots


```{r color_hist_facet}
penguins %>%
  ggplot() + aes(x = bill_l, fill = species) + 
  geom_histogram() +
  facet_wrap(~species) + 
  labs( x = 'Bill length in mm') +
  scale_fill_manual(values = wesanderson::wes_palette('Darjeeling1')) 
```



---

`r chunk_reveal("color_hist_facet", break_type = "auto")`

---

# Many other types of plots 

## Comparing plots - exercise


Change the previous graph to add a fitted density line

```{r color_hist_dens_ex}

penguins %>% ...
```


---


```{r color_hist_dens_cor}

 penguins %>%  ggplot() + aes(x = bill_l, y = ..density..) +
  facet_wrap(~species) + 
  geom_histogram(alpha=0.5, aes( fill = species)) +
   geom_density(aes(col = species)) + 
  labs( x = 'Bill length in mm') + #BREAK
  scale_fill_manual(values = wesanderson::wes_palette('Darjeeling1')) +  #BREAK
  scale_color_manual(values = wesanderson::wes_palette('Darjeeling1'))   

```

`r chunk_reveal("color_hist_dens_cor", break_type = "user")`


---


# Many other types of plots 

## Boxplot

```{r color_boxplot, eval = TRUE, echo = TRUE}
 penguins %>%  ggplot() + aes(x = species,  y = bill_l) +
  geom_boxplot(alpha=0.5, aes( fill = species)) +
  labs( y = 'Bill length in mm') + #BREAK
  scale_fill_manual(values = wesanderson::wes_palette('Darjeeling1')) +  #BREAK
  scale_color_manual(values = wesanderson::wes_palette('Darjeeling1'))   
```

---

```{r color_boxplot_2}
 penguins %>%  ggplot() + aes(x = species,  y = bill_l) +
  geom_violin(alpha=0.5, aes( fill = species)) +
  labs( y = 'Bill length in mm') + #BREAK
  scale_fill_manual(values = wesanderson::wes_palette('Darjeeling1')) 
```


`r chunk_reveal("color_boxplot_2", break_type = "user")`

---


```{r color_boxplot_3}
penguins %>%  ggplot() + aes(x = species,  y = bill_l) +
  geom_boxplot(alpha=0.5, aes( fill = species)) +
  geom_jitter(color="black", size=0.4, alpha=0.8) +
  labs( y = 'Bill length in mm') + #BREAK
  scale_fill_manual(values = wesanderson::wes_palette('Darjeeling1')) 
```

`r chunk_reveal("color_boxplot_3", break_type = "user")`

---


# Handling superscript and subscript


```{r color_boxplot_sup_sub, echo = TRUE, eval = TRUE}
penguins %>%  mutate(mu = bill_l * bill_d) %>% 
  ggplot() + aes(y= mu ) +
  geom_boxplot(alpha=0.5, aes( fill = species)) +
  labs( y = bquote(mu~(mm^2))) +
  scale_fill_manual(values = wesanderson::wes_palette('Darjeeling1')) 
```



---
name: useful_package 

# Useful packages for publication

The `ggpubr`  package is very helpful for publication.

---
template: useful_package
count: false

```{r ggpubr_p1, eval = TRUE, echo = TRUE}
gg_p1 <- gg
gg_p1
``` 


---
template: useful_package
count: false

```{r ggpubr_p2, eval = TRUE, echo = TRUE}
gg_p2 <- penguins %>%
  ggplot()  + aes(x = bill_l, y = ..density..) + geom_histogram(alpha=0.5, aes( fill = species)) +
   geom_density(aes(col = species)) +
  labs( x = 'Bill length in mm') +
  scale_fill_manual(values = wesanderson::wes_palette('Darjeeling1')) +
  scale_color_manual(values = wesanderson::wes_palette('Darjeeling1'))
gg_p2
```


---
template: useful_package
count: false

```{r ggpubr_p3, eval = TRUE, echo = TRUE}
##install.packages('ggpubr')
ggpubr::ggarrange(gg_p1, gg_p2, nrow=2, ncol = 1)
```

---
template: useful_package
count: false

```{r ggpubr_p4, eval = TRUE, echo = TRUE}
ggpubr::ggarrange(gg_p1, gg_p2, nrow=2, ncol = 1, common.legend = TRUE)
```

---
name: bivariate_explore 

# Useful packages for publications/explorations

The `GGally`  package contains plenty of very useful and nice functions

---
template: bivariate_explore
count: false

```{r ggpairs, eval = TRUE, echo = TRUE}
##install.packages('GGally')
library(GGally)
penguins %>% ggpairs(columns = c(1,3,4,5), mapping = aes(col = species)) +
scale_color_manual(values = wesanderson::wes_palette('Darjeeling1'))+
scale_fill_manual(values = wesanderson::wes_palette('Darjeeling1')) + theme(text = element_text(size = 6))
```

---
template: bivariate_explore
count: false

```{r ggpairs_fancy, eval = TRUE, echo = TRUE}
penguins %>% ggpairs(columns = c(1,3,4,5), mapping = aes(col = species),
                     upper = list(continuous = wrap( "cor",size = 2)),
                     lower = list(continuous = wrap('points', size = .5))) +
  scale_color_manual(values = wesanderson::wes_palette('Darjeeling1'))+
  scale_fill_manual(values = wesanderson::wes_palette('Darjeeling1')) + theme(text = element_text(size = 6))
```

---

name: gganimate

# Useful packages for presentation

The `gganimate`  package produces animated graph (html only).


```{r ani_final, eval = T, echo = T}
library(gganimate)
gg  +
  transition_states(year)  +
  geom_text(x = 56 , y = 15,
            aes(label = as.character(year)),
            size = 8, col = "grey50") +
  theme(legend.position="bottom") 
```


---
name: plotly

# Useful packages for online reporting

The `plotly`  package produces interactive plots  (html only).


```{r plotly, eval = T, echo = T, fig.height=5, fig.width=4}
library(plotly)
gg  %>%  ggplotly()
```

---

# Inspiring websites

- [R Graph Gallery](https://www.r-graph-gallery.com/)
- [Top50 ggplot Visualisation](http://r-statistics.co/Top50-Ggplot2-Visualizations-MasterList-R-Code.html)

---

# References


```{r refs, echo=FALSE, results="asis", eval = TRUE, cache = FALSE}
PrintBibliography(myBib)
```