ggplot2

background-image: url("img/logo_padded.001.jpeg")
background-position: left
background-size: 60%
class: middle, center,

<br>

## .base_color[Graphing with]

## .base_color[`ggplot2`]

<br>

### .navy[Kelly McConville]

#### .navy[ Stat 108 | Week 2 | Spring 2023]

]

---

## Announcements

* [Office Hours Schedule](https://docs.google.com/spreadsheets/d/1HqEmr4tEtFPWRrF5TJHd1VgtVD030w6aAYySoFTnhBw/edit?usp=share_link)
    + Some office hours will start in Week 3 so make sure to check the schedule.
* [Section Schedule](https://docs.google.com/spreadsheets/d/1DkR88j_jIkn9ZrA30FA8u3EAusAUaDN5TGhtuUVEyYA/edit?usp=share_link)
* Dates for Project 1 and Project 2 now in the [Tentative Schedule](https://docs.google.com/spreadsheets/d/1Ejyq-jcg7aCEqI3W4ahx4LrISTNewV2s6u-xozOx6Zg/edit?usp=sharing)
* Week 2 Lecture Quiz released at 10:30am on Wednesday.
    + Due at noon on Friday.
* P-Set 1 released at 9am on Thursday.
    + Will discuss how to access the p-sets on Wednesday.
    + Due at 5pm the following Wednesday

---

## Week 2 Goals

**Mon Lecture**

* Basics of `ggplot2`

* Explore several `geoms`.

* And a little data wrangling with `dplyr` as needed!

]

**Wed Lecture**

* Context!
    + Labels
    + Highlighting 
    + Useful text
* Look at more `geom`s.
* Explore further customizations.
    + Color
    + Themes
* Learn how to ask coding questions well.

]

---

### Data Viz Considerations

* We could spend all semester on data viz principles.

* Be thoughtful and iterate.

* Let's spend some time considering the strengths and weaknesses of the [graphs we encountered last time](https://docs.google.com/presentation/d/17cq7fA3noVyeTu-5syzY6e2Lq5rdSZo-n81hSbmrjFc/edit?usp=sharing).
    + You identified:
        + Clearest story
        + Most memorable
        + Best overall

* Spending time with well made (though not always perfect) graphs is a good way to get ideas.  Good sources:
    + [NYTimes: What's Going on in this Graph?]()
    + [https://fivethirtyeight.com/](https://fivethirtyeight.com/)
    + [Mona Chalabi's work](https://monachalabi.com)

---

## Our Voting Results

* Disclaimer: I teach Stat 100 right after Stat 108 and I think a Stat 100 student took off with half of our graphs!
    + Update: Found them...

---

## Recall: The Grammar of Graphics

* **data**: dataset that contains the data

* **geom**: geometric shape that the data are mapped to
    + point, line, bar, text, ...
    
* **aes**thetic: visual properties of the **geom** 
    + x position, y position, color, fill, shape
    
* **coord**: coordinate system
    + Cartesian, polar, geographic
    
* **scale**: controls how data are mapped to the visual values of the aesthetic
    + EX: particular colors, linear

* **guide**: legend to help user convert visual display back to the data

]

]

---

## ggplot2 example code

```r
ggplot(data = ---, mapping = aes(---)) +
  geom_---(---) + 
  coord_---() + 
  scale_---_---() +
  ---
```

---

### Example: Over the course of a year, how does the daily number of births vary?

* What patterns do you see?

---

### Example

```r
# Load library that has dataset of interest
library(mosaicData)

# Grab data
data(Births2015)

# Load tidyverse (which contains ggplot2)
library(tidyverse)
```

---

### Example

```r
# Example code
ggplot(data = ---, mapping = aes(---)) +
  geom_---(---) + 
  coord_---() + 
  scale_---_---() +
  ---
```

```r
# Create plot
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births)) + 
  geom_point()
```

]

]

---

### Example: Cycles related to day of the week?

```r
# Look at structure of data with dplyr function
glimpse(Births2015)
```

```
## Rows: 365
## Columns: 8
## $ date         <date> 2015-01-01, 2015-01-02, 2015-01-03, 2015-01-04, 2015-01-…
## $ births       <dbl> 8068, 10850, 8328, 7065, 11892, 12425, 12141, 12094, 1186…
## $ wday         <ord> Thu, Fri, Sat, Sun, Mon, Tue, Wed, Thu, Fri, Sat, Sun, Mo…
## $ year         <dbl> 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 201…
## $ month        <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
## $ day_of_year  <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17…
## $ day_of_month <dbl> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17…
## $ day_of_week  <dbl> 5, 6, 7, 1, 2, 3, 4, 5, 6, 7, 1, 2, 3, 4, 5, 6, 7, 1, 2, …
```

---

### Example: Cycles related to day of the week?

```r
# Create plot
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births, 
                     color = day_of_week)) + 
  geom_point()
```

* Additional aesthetic:
    + Why is this not quite what we want? What do we want?
* What happened to the aspect ratio when we added the legend?

]

]

---

### Example: Cycles related to day of the week?

```r
# Create plot
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births, 
                     color = wday)) + 
  geom_point() + 
  theme(legend.position = "bottom")
```

* You can also adjust the aspect ratio with the R chunk option: `fig.asp`

* What if we want to see the *direction* that the number of births take over time for each day of the week?
    + New visual cue/`geom`?

]

]

---

### Example

```r
# Create plot
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births, 
                     color = wday)) + 
  geom_line() + 
  theme(legend.position = "bottom")
```

* What if we want visual cues for both **position** and **direction**?

]

]

---

### Example

```r
# Create plot
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births, 
                     color = wday)) + 
  geom_line() +
  geom_point() +
  theme(legend.position = "bottom")
```

]

]

---

### Coordinate System Layer

```r
library(lubridate)

ggplot(data = Births2015, 
       mapping = aes(x = date, y = births, 
                     color = wday)) + 
  geom_line() + 
  geom_point() +
  theme(legend.position = "bottom") +
  coord_cartesian(xlim = 
                    as_date(c("2015-01-01",
                              "2015-01-31")))
```

* How did this new layer change our plot?

]

]

---

### Setting instead of Mapping

```r
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births, 
                     color = "midnightblue")) + 
  geom_point() 
```

* What if we want all the points to be colored "midnightblue"?

]

]

---

### Setting instead of Mapping

```r
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births)) + 
  geom_point(color = "midnightblue") 
```

* Here we'd say:
    + `date` is **mapped** to `x`
    + `births` is **mapped** to `y`
    + "midnightblue" **set** the `color`

]

]

---

### Setting instead of Mapping

```r
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births)) + 
  geom_point(color = "#ff006e") 
```

* Can also supply hex codes for colors!

]

]

---

### Setting instead of Mapping

```r
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births,
                     color = wday)) + 
  geom_point(color = "#ff006e") + 
  geom_line() + 
  theme(legend.position = "bottom")
```

* Difference between setting/mapping aesthetics in the base layer or a specific `geom` layer

]

]

---

### Layer order (sometimes) matters

```r
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births,
                     color = wday)) + 
  geom_line() + 
  geom_point(color = "#ff006e") + 
  theme(legend.position = "bottom")
```

]

]

---

### Let's explore other **geom**s

* Many are listed on the first page of the [`ggplot2` cheatsheet](https://raw.githubusercontent.com/rstudio/cheatsheets/main/data-visualization.pdf).

* Can also ask R:

```r
apropos("geom_")
```

```
##  [1] "geom_abline"            "geom_area"              "geom_bar"              
##  [4] "geom_bin_2d"            "geom_bin2d"             "geom_blank"            
##  [7] "geom_boxplot"           "geom_col"               "geom_contour"          
## [10] "geom_contour_filled"    "geom_count"             "geom_crossbar"         
## [13] "geom_curve"             "geom_density"           "geom_density_2d"       
## [16] "geom_density_2d_filled" "geom_density2d"         "geom_density2d_filled" 
## [19] "geom_dotplot"           "geom_errorbar"          "geom_errorbarh"        
## [22] "geom_freqpoly"          "geom_function"          "geom_hex"              
## [25] "geom_histogram"         "geom_hline"             "geom_jitter"           
## [28] "geom_label"             "geom_line"              "geom_linerange"        
## [31] "geom_map"               "geom_path"              "geom_point"            
## [34] "geom_pointrange"        "geom_polygon"           "geom_qq"               
## [37] "geom_qq_line"           "geom_quantile"          "geom_raster"           
## [40] "geom_rect"              "geom_ribbon"            "geom_rug"              
## [43] "geom_segment"           "geom_sf"                "geom_sf_label"         
## [46] "geom_sf_text"           "geom_smooth"            "geom_spoke"            
## [49] "geom_step"              "geom_text"              "geom_tile"             
## [52] "geom_violin"            "geom_vline"             "update_geom_defaults"
```

---

### Adding Curve(s)

```r
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births,
                     color = wday)) + 
  geom_point() + 
  geom_smooth(method = "lm", se = FALSE) + 
  theme(legend.position = "bottom")
```

* Does a multiple linear regression line(s) capture the trend?

]

]

---

### Adding Curve(s)

```r
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births,
                     color = wday)) + 
  geom_point() + 
  geom_smooth(se = FALSE) + 
  theme(legend.position = "bottom")
```

* The default LOESS smoother usually does a reasonable job.

]

]

---

### Adding Curve(s)

```r
ggplot(data = Births2015, 
       mapping = aes(x = date, y = births,
                     color = wday)) + 
  geom_point() + 
  geom_smooth(color = "black", se = FALSE) +
  theme(legend.position = "bottom")
```

* What happened?

]

]

---

### New Example: Movies and the Bechdel Test

* Need a new dataset with more categorical variables

* **The Alison Bechdel Rule**: A movie passes the test if:
    + There are at least two named women in the picture
    + They have a conversation with each other at some point
    + That conversation isn’t about a male character

* Movies from 1970 - 2013

```r
movies <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-03-09/movies.csv') %>%
  filter(rated %in% c("R", "PG-13", "PG", "G"))
```

---

### New Example: Movies and the Bechdel Test

```r
glimpse(movies)
```

```
## Rows: 1,549
## Columns: 34
## $ year          <dbl> 2013, 2013, 2013, 2013, 2013, 2013, 2013, 2013, 2013, 20…
## $ imdb          <chr> "tt2024544", "tt1272878", "tt0453562", "tt1335975", "tt1…
## $ title         <chr> "12 Years a Slave", "2 Guns", "42", "47 Ronin", "A Good …
## $ test          <chr> "notalk-disagree", "notalk", "men", "men", "notalk", "ok…
## $ clean_test    <chr> "notalk", "notalk", "men", "men", "notalk", "ok", "ok", …
## $ binary        <chr> "FAIL", "FAIL", "FAIL", "FAIL", "FAIL", "PASS", "PASS", …
## $ budget        <dbl> 2.00e+07, 6.10e+07, 4.00e+07, 2.25e+08, 9.20e+07, 1.20e+…
## $ domgross      <chr> "53107035", "75612460", "95020213", "38362475", "6734919…
## $ intgross      <chr> "158607035", "132493015", "95020213", "145803842", "3042…
## $ code          <chr> "2013FAIL", "2013FAIL", "2013FAIL", "2013FAIL", "2013FAI…
## $ budget_2013   <dbl> 2.00e+07, 6.10e+07, 4.00e+07, 2.25e+08, 9.20e+07, 1.20e+…
## $ domgross_2013 <chr> "53107035", "75612460", "95020213", "38362475", "6734919…
## $ intgross_2013 <chr> "158607035", "132493015", "95020213", "145803842", "3042…
## $ period_code   <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ decade_code   <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ imdb_id       <chr> "2024544", "1272878", "0453562", "1335975", "1606378", "…
## $ plot          <chr> "In the antebellum United States, Solomon Northup, a fre…
## $ rated         <chr> "R", "R", "PG-13", "PG-13", "R", "R", "PG-13", "PG-13", …
## $ response      <lgl> TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TR…
## $ language      <chr> "English", "English, Spanish", "English", "English, Japa…
## $ country       <chr> "USA, UK", "USA", "USA", "USA", "USA", "UK", "USA", "USA…
## $ writer        <chr> "John Ridley (screenplay), Solomon Northup (based on \"T…
## $ metascore     <dbl> 97, 55, 62, 29, 28, 55, 48, 33, 90, 58, 52, 78, 83, 53, …
## $ imdb_rating   <dbl> 8.3, 6.8, 7.6, 6.6, 5.4, 7.8, 5.7, 5.0, 7.5, 7.4, 6.2, 7…
## $ director      <chr> "Steve McQueen", "Baltasar Kormákur", "Brian Helgeland",…
## $ released      <chr> "08 Nov 2013", "02 Aug 2013", "12 Apr 2013", "25 Dec 201…
## $ actors        <chr> "Chiwetel Ejiofor, Dwight Henry, Dickie Gravois, Bryan B…
## $ genre         <chr> "Biography, Drama, History", "Action, Comedy, Crime", "B…
## $ awards        <chr> "Won 3 Oscars. Another 131 wins & 137 nominations.", "1 …
## $ runtime       <chr> "134 min", "109 min", "128 min", "118 min", "98 min", "1…
## $ type          <chr> "movie", "movie", "movie", "movie", "movie", "movie", "m…
## $ poster        <chr> "http://ia.media-imdb.com/images/M/MV5BMjExMTEzODkyN15BM…
## $ imdb_votes    <dbl> 143446, 87301, 43608, 25735, 123837, 85871, 18973, 10826…
## $ error         <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
```

---

## What are useful `geom`s for describing amounts/frequencies?

---

### Amounts: `geom_bar`

```r
ggplot(data = movies,
       mapping = aes(x = binary)) + 
  geom_bar()
```

* Verbalize the mapping of **data** to **geom_bar()**. 
    + How is this mapping different from **geom_point()**?

]

]

---

### Another option: `geom_col`

```r
# First wrangle with dplyr
movies_ag <- count(movies, binary) 
movies_ag
```

```
## # A tibble: 2 × 2
##   binary     n
##   <chr>  <int>
## 1 FAIL     863
## 2 PASS     686
```

```r
ggplot(data = movies_ag, 
       mapping = aes(x = binary, y = n)) + 
  geom_col()
```

]

]

---

### `geom_point` again

```r
ggplot(data = movies_ag, 
       mapping = aes(x = binary, y = n)) + 
  geom_point(size = 4)
```

* If you are worried about the data-ink ratio...

]

]

---

### `geom_point` + `geom_segment`

```r
ggplot(data = movies_ag, 
       mapping = aes(x = binary, y = n)) + 
  geom_segment(mapping = aes(xend = binary), 
               yend = 0) + 
  geom_point(size = 10, color = "orange") +
  ylim(c(0, 875))
```

* Lollipop chart: compromise?

]

]

---

### Two categorical variables: `geom_bar`

```r
ggplot(data = movies,
       mapping = aes(x = rated,
                     fill = binary)) + 
  geom_bar()
```

* Describe the mapping.

]

]

---

### Two categorical variables: `geom_bar`

```r
ggplot(data = movies,
       mapping = aes(x = rated,
                     fill = binary)) + 
  geom_bar(position = "fill")
```

* Describe the mapping.

]

]

---

### Two categorical variables: `geom_bar`

```r
ggplot(data = movies,
       mapping = aes(x = rated,
                     fill = binary)) + 
  geom_bar(position = "dodge")
```

* Describe the mapping.

]

]

---

### Two categorical variables: `geom_tile`

```r
movies_ag <- count(movies, rated, binary)

ggplot(data = movies_ag,
       mapping = aes(x = rated,
                     y = binary,
                     fill = n)) + 
  geom_tile()
```

* Describe the mapping.

]

]

---

### Two categorical variables: `geom_tile`

```r
movies_ag <- count(movies, rated, binary)

ggplot(data = movies_ag,
       mapping = aes(x = rated,
                     y = binary,
                     fill = n)) + 
  geom_tile() +
  scale_fill_viridis_c(direction = -1)
```

* Change the `fill` scale!

]

]

---

### Can display more than frequencies!

```r
movies_ag <- group_by(movies,
                      rated, binary) %>%
  summarize(mean_budget = mean(budget))

ggplot(data = movies_ag,
       mapping = aes(x = rated,
                     y = binary,
                     fill = mean_budget)) + 
  geom_tile() +
  scale_fill_viridis_c(direction = -1) +
  theme(legend.position = "bottom")
```

]

]

---

## What are useful `geom`s (graphs) for visualizing distributions?

---
    
### Distributions: `geom_histogram`

```r
ggplot(movies, aes(x = budget)) +
  geom_histogram()
```

* Verbalize the mapping.

]

]

---

### Distributions: `geom_histogram`

```r
ggplot(movies, aes(x = budget)) +
  geom_histogram(bins = 50, 
                 color = "white",
                 fill = "darkcyan")
```

* Can modify the mapping via the `binwidth` or `bins` arguments

]

]

---
    
### Distributions: `geom_histogram`

```r
ggplot(movies, aes(x = budget, 
                   fill = binary)) +
  geom_histogram(bins = 50, 
                 color = "white")
```

* What is problematic about this graph?

]

]

---

### Distributions: `geom_histogram`

```r
ggplot(movies, aes(x = budget, 
                   fill = binary)) +
  geom_histogram(bins = 50,
                 alpha = 0.4,
                 position = "identity")
```

* Still problematic

]

]

---

### One option: Facetting

```r
ggplot(movies, aes(x = budget, 
                   fill = binary)) +
  geom_histogram(bins = 50) +
  facet_wrap(~binary) +
  guides(fill = "none")
```

]

]

---

### One option: Facetting

```r
ggplot(movies, aes(x = budget, 
                   fill = binary)) +
  geom_histogram() +
  facet_grid(rated ~ binary, scales = "free_y") +
  guides(fill = "none")
```

]

]

---

### Another option: `geom_density`

```r
ggplot(movies, aes(x = budget, 
                   fill = binary)) +
  geom_density(alpha = 0.4)  +
  theme(legend.position = "bottom")
```

]

]

---

### Distributions: `geom_density`

```r
ggplot(movies, aes(x = budget, 
                   fill = binary)) +
  geom_density(position = "fill") +
  theme(legend.position = "bottom")
```

* How did the mapping change?  What does this add to the story?

]

]

---

### Distributions: `geom_boxplot`

```r
ggplot(movies, aes(x = binary, 
                   y = budget)) +
  geom_boxplot()
```

]

]

---

### Distributions: `geom_boxplot`

```r
ggplot(movies, aes(x = binary, 
                   y = budget)) +
  geom_boxplot(varwidth = TRUE,
               notch = TRUE)
```

* What does `varwidth` do?
* Why might we add `notch = TRUE`?

]

]

---

### Distributions: `geom_boxplot`

```r
ggplot(movies, aes(x = binary, 
                   y = budget,
                   fill = rated)) +
  geom_boxplot()
```

]

]

---

### Distributions: `geom_violin`

```r
ggplot(movies, aes(x = binary, 
                   y = budget)) +
  geom_violin()
```

* Utility of the violin over the box?

]

]

---

### Distributions: `geom_violin`

```r
ggplot(movies, aes(x = binary, 
                   y = budget)) +
  geom_violin() + 
  geom_jitter(alpha = .1,
              width = .1,
              color = "darkcyan")
```

]

]

---

### Reminders