Data Viz with ggplot2

class: center, middle

# Graphing with `ggplot2`
<img src="img/hero_wall_pink.png" width="800px"/>

## Kelly McConville

.large[Math 241 | Week 2 | Spring 2021]

---

# Announcements

* Sign up at [GitHub](https://github.com/) (it is free) and enter your GitHub username into [this spreadsheet](https://docs.google.com/spreadsheets/d/1nvM8jJDUvp8H5iYF59aKNdnJlqy_aW18R6d6DBUPeSU/edit?usp=sharing) by end of day **today**.
    + Link can also be found in our class Slack channel.

* Make sure [to sign-in](https://docs.google.com/document/d/1QMXSF9TxsXj3j8M42mwTatGnB_-GayYvGmHWnnXd-sg/edit?usp=sharing).

* Week 2 readings posted on the [Schedule](https://reed-statistics.github.io/math241s21/schedule.html).

* Lab 1 due on Gradescope on Monday!
    + Colored pencils outside my office (end of 3rd floor of Library).

> "The use of statistics in journalism, like the use of statistics in general, will always involve artistry." -- Jonathan Stray

---

## ["One Dataset, Visualized 25 Ways"](https://flowingdata.com/2017/01/24/one-dataset-visualized-25-ways/#jp-carousel-47350)

* There are many ways to visualize a dataset.
    + What geometric object do I use?
    + How do I map the variables to the aesthetics?
    + What context should I provide?
    + Should I include animation? Interactivity?

* Questions:
    + Which graphs are most effective at telling a story?
    + Any problematic graphs?

* Important messages
    + Try many graphs before you settle on one.
    + Some of these are the **start** of an excellent graph.  
        + Revise, revise, revise!

---

## Goals for Today

* Review `ggplot2`.

* Continue to consider best practices for data viz.

---

## Components of Data Graphics

* **data**: dataset that contains the raw 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
    
* **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?

* Cycles?

---

#### Example

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

# Grab data
data(Births2015)

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

---

#### Example

```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
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, 118…
## $ wday         <ord> Thu, Fri, Sat, Sun, Mon, Tue, Wed, Thu, Fri, Sat, Sun, M…
## $ year         <dbl> 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 20…
## $ 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, 1…
## $ day_of_month <dbl> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 1…
## $ 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 what we want? What do we want?

---

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

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

* What happened to the aspect ratio when we added the legend?

---

#### Example

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

---

#### Example

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

* What if we want both *position* and *direction*?

---

#### Example

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

---

### Coordinate System Layer

```r
library(lubridate)
ggplot(data = Births2015, mapping = aes(x = date, y = births)) + 
  geom_point() + 
  coord_cartesian(xlim = as_date(c("2015-01-01","2015-01-31")))
```

* What did this do?

---

### Setting instead of Mapping

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

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

```
## Error in FUN(X[[i]], ...): object 'midnightblue' not found
```

---

### Setting instead of Mapping

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

---

### Setting instead of Mapping

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

---

### Setting instead of Mapping

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

---

### Setting instead of Mapping

* Consider the order of your ggplot layers!

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

---

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

* Many are listed on the first page of the `ggplot2` cheatsheet.

* Can also ask R:

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

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

---

### Adding Curve(s)

```r
ggplot(data = Births78, mapping = aes(x = date, y = births, 
                                      color = wday)) + 
  geom_point() + 
  geom_smooth(se = FALSE)
```

---

## Adding Curve(s)

```r
ggplot(data = Births78, mapping = aes(x = date, y = births, 
                                      color = wday)) + 
  geom_point() + 
  geom_smooth(color = "black", se = FALSE)
```

* What happened?

---

### New Example: HELPrct

* Need a new dataset
    + `HELPrct`: Health Evaluation and Linkage to Primary Care Randomized Clinical Trial
        + Subjects admitted for treatment for addiction to one of three substances.

```r
# Library with data
library(mosaicData)

# Grab data
data(HELPrct)
```

---

### HELPrct

```r
glimpse(HELPrct)
```

```
## Rows: 453
## Columns: 30
## $ age              <int> 37, 37, 26, 39, 32, 47, 49, 28, 50, 39, 34, 58, 58, …
## $ anysubstatus     <int> 1, 1, 1, 1, 1, 1, NA, 1, 1, 1, NA, 0, 1, 1, 1, 1, 1,…
## $ anysub           <fct> yes, yes, yes, yes, yes, yes, NA, yes, yes, yes, NA,…
## $ cesd             <int> 49, 30, 39, 15, 39, 6, 52, 32, 50, 46, 46, 49, 22, 3…
## $ d1               <int> 3, 22, 0, 2, 12, 1, 14, 1, 14, 4, 0, 3, 5, 10, 2, 6,…
## $ daysanysub       <int> 177, 2, 3, 189, 2, 31, NA, 47, 31, 115, NA, 192, 6, …
## $ dayslink         <int> 225, NA, 365, 343, 57, 365, 334, 365, 365, 382, 365,…
## $ drugrisk         <int> 0, 0, 20, 0, 0, 0, 0, 7, 18, 20, 8, 0, 0, 0, 0, 0, 0…
## $ e2b              <int> NA, NA, NA, 1, 1, NA, 1, 8, 7, 3, NA, NA, NA, 1, NA,…
## $ female           <int> 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0…
## $ sex              <fct> male, male, male, female, male, female, female, male…
## $ g1b              <fct> yes, yes, no, no, no, no, yes, yes, no, no, no, no, …
## $ homeless         <fct> housed, homeless, housed, housed, homeless, housed, …
## $ i1               <int> 13, 56, 0, 5, 10, 4, 13, 12, 71, 20, 0, 13, 20, 13, …
## $ i2               <int> 26, 62, 0, 5, 13, 4, 20, 24, 129, 27, 0, 13, 31, 20,…
## $ id               <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 1…
## $ indtot           <int> 39, 43, 41, 28, 38, 29, 38, 44, 44, 44, 34, 11, 40, …
## $ linkstatus       <int> 1, NA, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, …
## $ link             <fct> yes, NA, no, no, yes, no, no, no, no, no, no, no, no…
## $ mcs              <dbl> 25.1, 26.7, 6.8, 44.0, 21.7, 55.5, 21.8, 9.2, 22.0, …
## $ pcs              <dbl> 58, 36, 75, 62, 37, 46, 25, 65, 38, 23, 60, 42, 39, …
## $ pss_fr           <int> 0, 1, 13, 11, 10, 5, 1, 4, 5, 0, 0, 13, 13, 1, 1, 7,…
## $ racegrp          <fct> black, white, black, white, black, black, black, whi…
## $ satreat          <fct> no, no, no, yes, no, no, yes, yes, no, yes, no, yes,…
## $ sexrisk          <int> 4, 7, 2, 4, 6, 5, 8, 6, 8, 0, 2, 0, 1, 4, 8, 3, 4, 4…
## $ substance        <fct> cocaine, alcohol, heroin, heroin, cocaine, cocaine, …
## $ treat            <fct> yes, yes, no, no, no, yes, no, yes, no, yes, yes, no…
## $ avg_drinks       <int> 13, 56, 0, 5, 10, 4, 13, 12, 71, 20, 0, 13, 20, 13, …
## $ max_drinks       <int> 26, 62, 0, 5, 13, 4, 20, 24, 129, 27, 0, 13, 31, 20,…
## $ hospitalizations <int> 3, 22, 0, 2, 12, 1, 14, 1, 14, 4, 0, 3, 5, 10, 2, 6,…
```

---

### Amounts

* What is a common geom for size/amount?

---

### `geom_bar`

```r
ggplot(data = HELPrct, mapping = aes(x = substance)) + 
  geom_bar()
```

* Aesthetic?

* Verbalize the mapping of **data** to **geom**. 
    + How is this mapping different from the scatterplot?

---

### Another option

```r
# First wrangle with dplyr
HELPrct %>%
  count(substance) %>%
ggplot(mapping = aes(x = substance, y = n)) + 
  geom_col()
```

* What changed?

---

### `geom_point`: New interpretation

```r
HELPrct %>%
  count(substance) %>%
ggplot(mapping = aes(x = substance, y = n)) + 
  geom_point(size = 4)
```

<img src="slidesWk2Tu_files/figure-html/unnamed-chunk-25-1.png" width="504" />
   
* Pros/Cons of `geom_point` versus `geom_bar`?

---

### `geom_point`: New interpretation

```r
HELPrct %>%
  count(substance) %>%
ggplot(mapping = aes(x = substance, y = n)) +
  geom_segment(mapping = aes(xend = substance), 
               yend = 0) + 
  geom_point(size = 4, color = "orange") 
```

<img src="slidesWk2Tu_files/figure-html/unnamed-chunk-26-1.png" width="504" />
   
* Pros/Cons of `geom_point` versus `geom_bar`?

---

### `geom_bar`: Two variables

* Relationship between substance and racegrp?

```r
ggplot(data = HELPrct, mapping = aes(x = substance, 
                                     fill = racegrp)) + 
  geom_bar()
```

* Describe the mapping.

---

### `geom_bar`: Two variables

* Relationship between substance and racegrp?

```r
ggplot(data = HELPrct, mapping = aes(x = substance, 
                                     fill = racegrp)) + 
  geom_bar(position = "fill")
```

* Describe the mapping.

---

### `geom_bar`: Two variables

* Relationship between substance and racegrp?

```r
ggplot(data = HELPrct, mapping = aes(x = substance, 
                                     fill = racegrp)) + 
  geom_bar(position = "dodge")
```

<img src="slidesWk2Tu_files/figure-html/unnamed-chunk-29-1.png" width="504" />
  
* Describe the mapping.

---

### `geom_tile`: Two variables

* Relationship between substance and racegrp?

```r
HELPrct %>%
  count(substance, racegrp) %>%
ggplot(mapping = aes(x = substance, y = racegrp,
                     fill = n)) + 
  geom_tile()
```

<img src="slidesWk2Tu_files/figure-html/unnamed-chunk-30-1.png" width="504" />
  
* Describe the mapping.

---

### Distributions

* What are useful graphs/geoms for visualizing distributions?

---

### `geom_histogram`

* How old are participants in the study?

```r
ggplot(HELPrct, aes(x = age)) +
  geom_histogram()
```

* Verbalize the mapping.

---

### `geom_histogram`

* Can modify the mapping via the `binwidth` argument

```r
ggplot(HELPrct, aes(x = age)) +
  geom_histogram(binwidth = 2)
```

---

### `geom_histogram`

* Does the age distribution vary by substance?

```r
ggplot(HELPrct, aes(x = age, fill = substance)) +
  geom_histogram()
```

* Bad

---

### `geom_histogram`

* Does the age distribution vary by substance?

```r
ggplot(HELPrct, aes(x = age, fill = substance)) +
  geom_histogram(alpha = 0.4, position = "identity")
```

* Still bad

---

### Facetting

* Does the age distribution vary by substance?

```r
ggplot(HELPrct, aes(x = age)) +
  geom_histogram() +
  facet_wrap(~substance)
```

---

### `geom_density`

```r
ggplot(HELPrct, aes(x = age)) +
  geom_density()
```

---

### `geom_density`

* Does the age distribution vary by substance?

```r
ggplot(HELPrct, aes(x = age, color = substance)) +
  geom_density()
```

---

### `geom_density`

* Does the age distribution vary by substance?

```r
ggplot(HELPrct, aes(x = age, fill = substance)) +
  geom_density(alpha = 0.3)
```

---

### `geom_density`

* Does the age vary by substance?

```r
ggplot(HELPrct, aes(x = age, fill = substance)) +
  geom_density(position = "fill", adjust = 2)
```

---

### Faceting

* Does the age vary by substance?

```r
ggplot(HELPrct, aes(x = age)) +
  geom_line(stat = "density") +
  facet_grid(.~substance)
```

---

### `geom_boxplot`

* Does the age distribution vary by substance?

```r
ggplot(HELPrct, aes(x = substance, y = age)) +
  geom_boxplot()
```

---

### `geom_boxplot`

* Does the age distribution vary by substance?

```r
ggplot(HELPrct, aes(x = substance, y = age)) +
  geom_boxplot(varwidth = TRUE)
```

* What does `varwidth` do?

---

### `geom_boxplot`

* Does the age distribution vary by substance?

```r
ggplot(HELPrct, aes(x = substance, y = age)) +
  geom_boxplot(notch = TRUE)
```

* Why might we add `notch = TRUE`?

---

### `geom_boxplot`

* Does the age distribution vary by substance and racegrp?

```r
ggplot(HELPrct, aes(x = substance, y = age, 
                    color = racegrp)) +
  geom_boxplot()
```

---

### `geom_boxplot`

* Does the age distribution vary by substance and racegrp?

```r
ggplot(HELPrct, aes(x = substance, y = age, 
                    color = racegrp)) +
  geom_boxplot() + 
  coord_flip()
```

---

### `geom_violin`

* Does the age distribution vary by substance?
    + Want to see distribution more clearly!

```r
ggplot(HELPrct, aes(x = substance, y = age)) +
  geom_violin()
```

* Utilty of the violin over the box?

---

### `geom_violin`

* Does the age distribution vary by substance?
    + Want to see distribution more clearly!

```r
ggplot(HELPrct, aes(x = substance, y = age)) +
  geom_violin() + 
  geom_jitter(alpha = .3, width = .1)
```

---

## Thursday

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