class: center, middle

# Data Visualization: `ggplot2`

<span style="color: #91204D;"> .large[Kelly McConville | Math 141 | Week 2 | Fall 2020] </span>

---

## Announcements

* Lab 01 is due BEFORE your Week 2 lab (regardless of what Gradescope says). 
    + Questions on uploading to Gradescope?
    + Questions on exporting the pdf from RStudio?
    
* Don't forget to come by office hours twice during the first four weeks of the semester!

---

## Week 2 Topics

* **Creating** Data Visualizations

---

# Goals for Today

* Discuss standard graphs for numerical data: 
    + **Histogram**: one numerical variable
    + **Side-by-side boxplot**: one numerical variable and one categorical variable
    + **Side-by-side violin plot**: one numerical variable and one categorical variable
    + **Scatterplot**: two numerical variables
    + **Linegraph**: two numerical variables

* Learn how to build these graphs with `ggplot2`.

--
    
* On Friday:
    + Bar plots for categorical data.
    + How to incorporate more than two variables into a plot.
    + Context!
        + Won't worry today about labs, titles, etc.

---

# Load packages

* `ggplot2` is part of this collection of data science packages.

```r
# Load necessary packages
library(tidyverse)
```

---

## Load the [Portland Biketown data](https://www.biketownpdx.com/system-data)

```r
# Import the data
biketown <- 
read_csv("/home/courses/math141f20/Data/biketown_spring1920.csv")

# Look at the data
glimpse(biketown)
```

```
## Rows: 124,726
## Columns: 19
## $ RouteID          <dbl> 10889392, 10889395, 10889396, 10889399, 10889407, 10…
## $ PaymentPlan      <chr> "Casual", "Subscriber", "Subscriber", "Subscriber", …
## $ StartHub         <chr> "N Mississippi at Beech", NA, NA, "SW Moody at Aeria…
## $ StartLatitude    <dbl> 46, 45, 46, 45, 46, 46, 46, 46, 46, 46, 46, 46, 46, …
## $ StartLongitude   <dbl> -123, -123, -123, -123, -123, -123, -123, -123, -123…
## $ StartDate        <chr> "3/1/2019", "3/1/2019", "3/1/2019", "3/1/2019", "3/1…
## $ StartTime        <time> 00:22:00, 00:30:00, 00:33:00, 00:35:00, 00:46:00, 0…
## $ EndHub           <chr> "SE 29th at Stark", "NE Multnomah at Grand", NA, "NE…
## $ EndLatitude      <dbl> 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, 46, …
## $ EndLongitude     <dbl> -123, -123, -123, -123, -123, -123, -123, -123, -123…
## $ EndDate          <chr> "3/1/2019", "3/1/2019", "3/1/2019", "3/1/2019", "3/1…
## $ EndTime          <time> 00:48:00, 01:04:00, 01:03:00, 01:06:00, 00:55:00, 0…
## $ TripType         <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
## $ BikeID           <dbl> 6023, 7313, 6538, 6424, 6576, 7324, 6668, 7266, 7320…
## $ BikeName         <chr> "0323 BIKETOWN", "0985 DESIGN BIKE, N PAUL BUNYAN", …
## $ Distance_Miles   <dbl> 4.07, 3.61, 0.98, 3.28, 1.25, 2.70, 0.82, 0.11, 1.43…
## $ Duration         <time> 00:26:21, 00:34:18, 00:30:22, 00:30:58, 00:08:56, 0…
## $ RentalAccessPath <chr> "mobile", "keypad", "keypad", "keypad", "mobile", "k…
## $ MultipleRental   <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FAL…
```

---

```r
# Look at the data
head(biketown)
```

```
## # A tibble: 6 x 19
##   RouteID PaymentPlan StartHub StartLatitude StartLongitude StartDate StartTime
##     <dbl> <chr>       <chr>            <dbl>          <dbl> <chr>     <time>   
## 1  1.09e7 Casual      N Missi…          45.5          -123. 3/1/2019  22'00"   
## 2  1.09e7 Subscriber  <NA>              45.5          -123. 3/1/2019  30'00"   
## 3  1.09e7 Subscriber  <NA>              45.5          -123. 3/1/2019  33'00"   
## 4  1.09e7 Subscriber  SW Mood…          45.5          -123. 3/1/2019  35'00"   
## 5  1.09e7 Subscriber  NW 18th…          45.5          -123. 3/1/2019  46'00"   
## 6  1.09e7 Subscriber  <NA>              45.6          -123. 3/1/2019  53'00"   
## # … with 12 more variables: EndHub <chr>, EndLatitude <dbl>,
## #   EndLongitude <dbl>, EndDate <chr>, EndTime <time>, TripType <lgl>,
## #   BikeID <dbl>, BikeName <chr>, Distance_Miles <dbl>, Duration <time>,
## #   RentalAccessPath <chr>, MultipleRental <lgl>
```

---

```r
# To access one variable: dataset$variable
head(biketown$Distance_Miles)
```

```
## [1] 4.07 3.61 0.98 3.28 1.25 2.70
```

```r
# Determine type 
class(biketown$Distance_Miles)
```

```
## [1] "numeric"
```

```r
class(biketown)
```

```
## [1] "spec_tbl_df" "tbl_df"      "tbl"         "data.frame"
```

```r
# Variable names
names(biketown)
```

```
##  [1] "RouteID"          "PaymentPlan"      "StartHub"         "StartLatitude"   
##  [5] "StartLongitude"   "StartDate"        "StartTime"        "EndHub"          
##  [9] "EndLatitude"      "EndLongitude"     "EndDate"          "EndTime"         
## [13] "TripType"         "BikeID"           "BikeName"         "Distance_Miles"  
## [17] "Duration"         "RentalAccessPath" "MultipleRental"
```

---

# Biketown data

```r
# Remove suspect points
biketown <- filter(biketown, Distance_Miles < 1000)

# Display variables
select(biketown, Distance_Miles, StartTime, PaymentPlan, Duration)
```

```
## # A tibble: 124,721 x 4
##    Distance_Miles StartTime PaymentPlan Duration
##             <dbl> <time>    <chr>       <time>  
##  1           4.07 00:22     Casual      26'21"  
##  2           3.61 00:30     Subscriber  34'18"  
##  3           0.98 00:33     Subscriber  30'22"  
##  4           3.28 00:35     Subscriber  30'58"  
##  5           1.25 00:46     Subscriber  08'56"  
##  6           2.7  00:53     Subscriber  46'44"  
##  7           0.82 01:33     Subscriber  06'55"  
##  8           0.11 01:57     Subscriber  02'14"  
##  9           1.43 02:00     Subscriber  27'53"  
## 10           0.53 02:04     Casual      05'07"  
## # … with 124,711 more rows
```

---

# Grammar of Graphics

* **data**: data frame that contains the raw data
    + Variables

* **geom**: geometric shape that the data are mapped to.
    + point, line, bar, text, ...
    
* **aesthetic**: visual properties of the **geom** 
    + x position, y position, color, fill, shape

* **scale**: controls how data are mapped to the visual values of the aesthetic.
    + EX: particular colors

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

---

# `ggplot2` example code

**Guiding Principle**: We will map variables from the **data** to the **aes**thetic attributes (e.g. location, size, shape, color) of **geom**etric objects (e.g. points, lines, bars).

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

* There are other layers, such as `scales_---_---()` and `labs()`, but we will wait on those.

---

# Histograms

* Binned counts of data.

* Great for assessing shape.

---

# Data Shapes

* Shapes: 
    + Right skewed
    + Bell shaped and symmetric
    + Left skewed

---

# Histograms

```r
# Create histogram
ggplot(data = biketown, mapping = aes(x = Distance_Miles)) + 
  geom_histogram()
```

---

# Histograms

```r
# Create histogram
ggplot(data = biketown, mapping = aes(x = Distance_Miles)) + 
  geom_histogram(color = "white", fill = "blue", bins = 70)
```

* Also can set (instead of map) some of the aesthetics.  If not mapping to a variable, put directly in the `geom_histogram()`.
---

# Boxplots

* Five number summary: minimum, first quartile (Q1), median, third quartile (Q3 ), maximum
* Interquartile range (IQR) `$=$` Q3 `$-$` Q1
* Outliers: unusual points 
    + Boxplot defines unusual as being beyond `$1.5*IQR$` from `$Q1$` or `$Q3$`.
* Whiskers: reach out to the furtherest point that is NOT an outlier

---

# Boxplots

```r
ggplot(data = biketown, mapping = aes(x = PaymentPlan, 
                                      y = Distance_Miles)) +
  geom_boxplot()
```

---

# Boxplots

```r
ggplot(data = biketown, mapping = aes(x = PaymentPlan, 
                                      y = Distance_Miles)) + 
  geom_boxplot() +
  coord_cartesian(ylim = c(0,8))
```

---

# Violin plots

```r
ggplot(data = biketown, mapping = aes(x = PaymentPlan, 
                                      y = Distance_Miles)) + 
  geom_violin() +
  coord_cartesian(ylim = c(0,8))
```

* Provides shape.

---

### Scatterplots

* Explore relationships between numerical variables.
    + We will be especially interested in **linear** relationships.

---

### Scatterplots

```r
ggplot(data = biketown, mapping = aes(x = Duration, 
                                      y = Distance_Miles)) + 
  geom_point()
```

---

# Fixing Over-plotting

```r
ggplot(data = biketown, mapping = aes(x = Duration, 
                                      y = Distance_Miles)) + 
  geom_point(alpha = 0.1)
```

---

# Scatterplots

```r
#Summarize use by time
biketown2 <- count(biketown, StartTime)
biketown2
```

```
## # A tibble: 1,440 x 2
##    StartTime     n
##    <time>    <int>
##  1 00'00"       28
##  2 01'00"       38
##  3 02'00"       40
##  4 03'00"       26
##  5 04'00"       31
##  6 05'00"       33
##  7 06'00"       28
##  8 07'00"       41
##  9 08'00"       33
## 10 09'00"       35
## # … with 1,430 more rows
```

---

# Scatterplots

```r
ggplot(data = biketown2, mapping = aes(x = StartTime, y = n)) + 
  geom_point()
```

---

# Linegraphs

* Also called time series plot if time is represented on the x axis.

---

# Linegraphs

```r
ggplot(data = biketown2, mapping = aes(x = StartTime, y = n)) + 
  geom_line()
```

* Also called time series plot if time is represented on the x axis.

---

# Recap: `ggplot2`

```r
library(tidyverse)
ggplot(data = ---, mapping = aes(---)) +
  geom_---(---) 
```