Week 13: Data Visualization

.leftcol30[
<center>
<img src="https://github.com/emse-p4a-gwu/emse-p4a-gwu.github.io/raw/master/images/p4a_hex_sticker.png" width=250>
</center>
]
.rightcol70[
# Week 13: .fancy[Data Visualization]

### EMSE 4574: Intro to Programming for Analytics
### John Paul Helveston
### November 24, 2020
]

---

# Quiz 6

.leftcol[
- ### Go to `#classroom` channel in Slack for link
- ### Open up RStudio before you start - you'll probably want to use it.
]
.rightcol[
<center>
<img src="images/quiz_doge.png" width="400">
</center>
]

---
# Before we start

Make sure you have the "tidyverse" installed and loaded, and import these two data frames

```r
library(tidyverse)
library(here)

birds <- read_csv(here('data', 'wildlife_impacts.csv'))
bears <- read_csv(here('data', 'bear_killings.csv'))
```

(this is at the top of the notes.R file)

---
# The Challenger disaster

On January 28, 1986 the space shuttle Challenger exploded

---
# The Challenger disaster

NASA Engineers had the data on temperature & o-ring failure

---
class: center

## What NASA was shown

.leftcol60[
<img src="images/rockets_chart.png" width=600>
]
.rightcol40[.left[
<br><br><br><br><br><br><br><br>
Tufte, Edward R. (1997) _Visual Explanations: Images and Quantities, Evidence and Narrative_, Graphics Press, Cheshire, Connecticut.]]

---
class: center

# What NASA _should_ have been shown

.left[Tufte, Edward R. (1997) _Visual Explanations: Images and Quantities,<br> Evidence and Narrative_, Graphics Press, Cheshire, Connecticut.]

---
class: inverse, middle

# Week 13: .fancy[Data Visualization]

## 1. Plotting with Base R
## 2. Plotting with **ggplot2**
## 3. Tweaking your ggplot

---

# Week 13: .fancy[Data Visualization]

## 1. .orange[Plotting with Base R]
## 2. Plotting with **ggplot2**
## 3. Tweaking your ggplot

---
# Today's data:
## Bear attacks in North America

Explore the `bears` data frame:

```r
glimpse(bears)
head(bears)
```

---
class: center

## Two basic plots in R

.leftcol[
### Scatterplots
<img src="slides-13-data-visualization_files/figure-html/base_scatter-1.png" width="504" />
]
.rightcol[
### Histograms
<img src="slides-13-data-visualization_files/figure-html/base_hist-1.png" width="504" />
]

---
# Scatterplots with `plot()`

### Plot relationship between two variables

--
.leftcol[
General syntax:

```r
plot(x = x_vector, y = y_vector)
```
]

---
# Scatterplots with `plot()`

### Plot relationship between two variables

```r
plot(x = x_vector, y = y_vector)
```

Example:

```r
var1 <- seq(1, 5)
var2 <- 2*var1
plot(x = var1, y = var2)
```
]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-12-1.png" width="432" />
]

---
# Scatterplots with `plot()`

### `x` and `y` must have the same length!

```r
var2 <- var2[-1]
```
--

```r
length(var1) == length(var2)
```

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

```r
plot(x = var1, y = var2)
```

```
## Error in xy.coords(x, y, xlabel, ylabel, log): 'x' and 'y' lengths differ
```

---
# Scatterplots with `plot()`

### Plotting variables from a data frame:

--
.leftcol[
Plot `year` vs. `age`:

```r
plot(x = bears$year, y = bears$age)
```
]
--
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-16-1.png" width="504" />
]

---
# Making `plot()` pretty

```r
plot(x    = bears$year,
     y    = bears$age,
     col  = 'darkblue', # Point color
     pch  = 19, # Point shape
     main = "Age of victims over time",
     xlab = "Year",
     ylab = "Age of victim")
```
]]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-17-1.png" width="504" />
]

---
class: inverse

## Think pair share: `plot()`

Does the annual number of bird impacts appear to be changing over time?

Make a plot using the `birds` data frame to justify your answer

Hint: You may need to create a summary data frame to answer this question!

**Bonus points**: Make your plot pretty

---
# Histograms with `hist()`

### Plot the _distribution_ of a single variable

```r
hist(x = x_vector)
```
]

---
# Histograms with `hist()`

### Plot the _distribution_ of a single variable

```r
hist(x = x_vector)
```

Example:

```r
hist(bears$month)
```
]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-21-1.png" width="432" />
]

---
# Making `hist()` pretty

```r
hist(x      = bears$month,
     breaks = 12,
     col    = 'darkred',
     main   = "Bear killings by month",
     xlab   = "Month",
     ylab   = "Count")
```
]]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-22-1.png" width="432" />
]

---
class: inverse

## Think pair share: `hist()`

Make plots using the `birds` data frame to answer these questions

- Which months have the highest and lowest number of bird impacts in the dataset?
- Which aircrafts experience more impacts: 2-engine, 3-engine, or 4-engine?
- At what height do most impacts occur?

**Bonus points**: Make your plots pretty

---

# Week 13: .fancy[Data Visualization]

## 1. Plotting with Base R
## 2. .orange[Plotting with **ggplot2**]
## 3. Tweaking your ggplot

---
class: center
## Advanced figures with `ggplot2`

<center>
<img src="images/horst_monsters_ggplot2.png" width=600>
</center>Art by [Allison Horst](https://www.allisonhorst.com/)

---
.leftcol[
<img src="images/making_a_ggplot.jpeg" width=600>
]
.rightcol[
# "Grammar of Graphics"

Concept developed by Leland Wilkinson (1999)

**ggplot2** package developed by Hadley Wickham (2005)
]

---
# Making plot layers with ggplot2

<br>
### 1. The data (we'll use `bears`)
### 2. The aesthetic mapping (what goes on the axes?)
### 3. The geometries (points? bars? etc.)

---
# Layer 1: The data

The `ggplot()` function initializes the plot with whatever data you're using

```r
ggplot(data = bears)
```
]
.rightcol[.blackborder[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-24-1.png" width="504" />
]]

---
# Layer 2: The aesthetic mapping

The `aes()` function determines which variables will be _mapped_ to the geometries<br>(e.g. the axes)

```r
ggplot(data = bears,
       aes(x = year, y = age))
```
]
.rightcol[.blackborder[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-25-1.png" width="504" />
]]

---
# Layer 3: The geometries

Use `+` to add geometries (e.g. points)

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
*   geom_point()
```
]
.rightcol[.blackborder[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-26-1.png" width="504" />
]]

---
# Other common geometries

- `geom_point()`: scatter plots
- `geom_line()`: lines connecting data points
- `geom_col()`: bar charts
- `geom_boxplot()`: boxes for boxplots

---
# Scatterplots with `geom_point()`

Add points:

---
# Scatterplots with `geom_point()`

Change the color of all points:

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
*   geom_point(color = 'blue')
```
]
.rightcol[.blackborder[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-28-1.png" width="504" />
]]

---
# Scatterplots with `geom_point()`

Map the point color to a **variable**:

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
*   geom_point(aes(color = gender))
```
Note that `color = gender` is _inside_ `aes()`
]
.rightcol[.blackborder[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-29-1.png" width="504" />
]]

---
# Scatterplots with `geom_point()`

Adjust labels with `labs()` layer:

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point(aes(color = gender)) +
*   labs(x = "Year",
*        y = "Age",
*        title = "Bear victim age over time",
*        color = "Gender")
```
]
.rightcol[.blackborder[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-30-1.png" width="504" />
]]

---
class: inverse

## Think pair share: `geom_point()`

Use the `birds` data frame to create the following plots

.leftcol[
<img src="slides-13-data-visualization_files/figure-html/ggpoint_p1-1.png" width="504" />
]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/ggpoint_p2-1.png" width="504" />
]

---
class: inverse, center

# .fancy[Break]

---

## Make bar charts with `geom_col()`

With bar charts, you'll often need to create summary variables to plot

--
.leftcol[
Step 1: Summarize the data

```r
bear_months <- bears %>%
* count(month)
```

Step 2: Make the plot

```r
ggplot(bear_months) +
* geom_col(aes(x = month, y = n))
```
]
.rightcol[
Example: count of attacks by month
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-33-1.png" width="504" style="display: block; margin: auto;" />
]

---
## Make bar charts with `geom_col()`

Alternative approach: piping directly into ggplot
.leftcol[

```r
bears %>%
* count(month) %>%  # Pipe into ggplot
  ggplot() +
  geom_col(aes(x = month, y = n))
```
]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-34-1.png" width="504" style="display: block; margin: auto;" />
]

---
## Be careful with `geom_col()` vs. `geom_bar()`

Map both `x` and `y`

```r
bears %>%
* count(month) %>%
  ggplot() +
* geom_col(aes(x = month, y = n))
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-35-1.png" width="288" />
]
.rightcol[
### .center[`geom_bar()`]

Only map `x` (`y` is computed)

```r
bears %>%
  ggplot() +
* geom_bar(aes(x = month))
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-36-1.png" width="288" />
]

---
## Make bar charts with `geom_col()`

```r
bears %>%
  filter(!is.na(age)) %>%
  group_by(year) %>%
* summarise(meanAge = mean(age)) %>%
  ggplot() +
* geom_col(aes(x = year, y = meanAge))
```
]
.rightcol[.blackborder[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-37-1.png" width="504" />
]]

---
### Change bar width:  `width`
### Change bar color: `fill`
### Change bar outline: `color`

```r
bears %>%
  count(month) %>%
  ggplot() +
  geom_col(aes(x = month, y = n),
*          width = 0.7,
*          fill = "blue",
*          color = "red")
```
]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-38-1.png" width="504" style="display: block; margin: auto;" />
]

---
### Map the `fill` to `bearType`

```r
bears %>%
* count(month, bearType) %>%
  ggplot() +
  geom_col(aes(x = month, y = n,
*              fill = bearType))
```
]

Note that I had to summarize the count by both `month` and `bearType`

```r
bears %>%
  count(month, bearType)
```
]
.code60[

```
## # A tibble: 27 x 3
##    month bearType     n
##    <dbl> <chr>    <int>
##  1     1 Brown        1
##  2     1 Polar        2
##  3     2 Brown        1
##  4     3 Brown        1
##  5     4 Black        1
##  6     4 Brown        3
##  7     5 Black       15
##  8     5 Brown        2
##  9     5 Polar        1
## 10     6 Black       10
## # … with 17 more rows
```
]]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-41-1.png" width="504" style="display: block; margin: auto;" />
]

---
# "Factors" = Categorical variables

By default, R makes numeric variables _continuous_

```r
bears %>%
  count(month) %>%
  ggplot() +
  geom_col(aes(x = month, y = n)) 
```
**The variable `month` is a _number_**
]
.rightcol[.blackborder[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-43-1.png" width="504" style="display: block; margin: auto;" />
]]

---
# "Factors" = Categorical variables

You can make a continuous variable _categorical_ using `as.factor()`

```r
bears %>%
  count(month) %>%
  ggplot() +
* geom_col(aes(x = as.factor(month),
               y = n)) 
```
]

**The variable `month` is a _factor_**
]
.rightcol[.blackborder[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-44-1.png" width="504" />
]]

---
class: inverse

## Think pair share: `geom_col()`

Use the `bears` and `birds` data frame to create the following plots

.leftcol[
<img src="slides-13-data-visualization_files/figure-html/ggbar_p1-1.png" width="504" />
]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/ggbar_p2-1.png" width="504" />
]

---

# Week 13: .fancy[Data Visualization]

## 1. Plotting with Base R
## 2. Plotting with **ggplot2**
## 3. .orange[Tweaking your ggplot]

---
# Working with themes

Themes change _global_ features of your plot, like the background color, grid lines, etc.

--
.leftcol[

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point()
```
]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-46-1.png" width="504" />
]

---
# Working with themes

Themes change _global_ features of your plot, like the background color, grid lines, etc.

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point() +
*   theme_bw()
```
]
.rightcol[
<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-47-1.png" width="504" />
]

---
### Common themes

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point() +
*   theme_bw()
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-48-1.png" width="432" />
]
.rightcol[
`theme_minimal()`

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point() +
*   theme_minimal()
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-49-1.png" width="432" />
]

---
### Common themes

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point() +
*   theme_classic()
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-50-1.png" width="432" />
]
.rightcol[
`theme_void()`

```r
ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point() +
*   theme_void()
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-51-1.png" width="432" />
]

---
### Other themes: [**hrbrthemes**](https://github.com/hrbrmstr/hrbrthemes)

```r
*library(hrbrthemes)

ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point() +
*   theme_ipsum()
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-52-1.png" width="432" />
]
.rightcol[

```r
*library(hrbrthemes)

ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point() +
*   theme_ft_rc()
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-53-1.png" width="432" />
]

---
### Other themes: **ggthemes**

```r
*library(ggthemes)

ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point() +
*   theme_economist()
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-54-1.png" width="432" />
]
.rightcol[

```r
*library(ggthemes)

ggplot(data = bears,
       aes(x = year, y = age)) +
    geom_point() +
*   theme_economist_white()
```

<img src="slides-13-data-visualization_files/figure-html/unnamed-chunk-55-1.png" width="432" />
]

---
# Save figures with `ggsave()`

--
First, assign the plot to an object name:

```r
scatterPlot <- ggplot(data = bears) +
    geom_point(aes(x = year, y = age))
```

--
Then use `ggsave()` to save the plot:

```r
ggsave(filename = here('plots', 'scatterPlot.png'),
       plot   = scatterPlot,
       width  = 6, # inches
       height = 4)
```

---
class: inverse

# Extra practice 1

Use the `mtcars` data frame to create the following plots

.cols3[
<img src="slides-13-data-visualization_files/figure-html/mtcars_1-1.png" width="324" />
]
.cols3[
<img src="slides-13-data-visualization_files/figure-html/mtcars_2-1.png" width="324" />
]
.cols3[
<img src="slides-13-data-visualization_files/figure-html/mtcars_3-1.png" width="324" />
]

---
class: inverse

# Extra practice 2

Use the `mpg` data frame to create the following plot