CS&SS 569 Visualizing Data and Models

.title[
# CS&SS 569 Visualizing Data and Models
]
.subtitle[
## Lab 3: ggplot2
]
.author[
### Brian Leung
]
.institute[
### Department of Political Science, UW
]
.date[
### 2024-01-19
]

---

## Introduction

- Unpack the design philosophy and inner working of ggplot2

- Gain finer control over *each and every* aspect of the graphic design, rather than relying on default

- Showcase some tricks and useful functions common in my own viz workflow

- Hopefully it benefits beginners, and intermediate (at least marginally), users to ggplot2

---

## ggplot2 design philosophy

- Grammar of graphics

- A statistical graphic is a `mapping` of `data` variables to `aes`thetic attributes of `geom`etric objects. (Wilkinson 2005)

- *What* data do you want to visualize?
  
--

- `ggplot(data = ...)`
  
--
  
- *How* are variables mapped to specific aesthetic attributes?

- `aes(... = ...)`
  
--

- positions (`x`, `y`), `shape`, `colour`, `size`, `fill`, `alpha`, `linetype`, `label`... 
  
--

- If the value of an attribute do not vary w.r.t. some variable, don't wrap it within `aes(...)`
  
--

- *Which* geometric shapes do you use to represent the data?

- `geom_{...}`: `geom_point`, `geom_line`, `geom_ribbon`, `geom_polygon`...
  
---

## ggplot2: A layered grammar

- ggplot2: A *layered* grammar of graphics (Wickham 2009)

--
  
  - Build a graphic from multiple layers; each consists of some geometric objects or transformation 
  
--

- Use `+` to stack up layers

- When you declare `ggplot(data = ..., aes(...))`, you're declaring the default specification

- Within each `geom_{}` layer that follows, two things are *inherited* from the default specification at the master level

- Data: inherited from the default data you supply

--
  
  - Aesthetics: inherited from the default aesthetics (`inherit.aes = TRUE`)

--
  
  - They are convenient but can create unintended consequences

- We will learn how to overwrite this property for extra flexibility

---

## Quick demo: inherit.aes = TRUE property

```r
library(tidyverse)

iver_data <- read_csv("https://faculty.washington.edu/cadolph/vis/iverRevised.csv")

head(iver_data)
```

```
# A tibble: 6 × 4
  country   povertyReduction effectiveParties partySystem 
  <chr>                <dbl>            <dbl> <chr>       
1 Australia             42.2             2.38 Majoritarian
2 Belgium               78.8             7.01 Proportional
3 Canada                29.9             1.69 Majoritarian
4 Denmark               71.5             5.04 Proportional
5 Finland               69.1             5.14 Proportional
6 France                57.9             2.68 Majoritarian
```

---

count: false
 
## Why are there multiple smooth curves? -- inherit.aes
.panel1-iver1-non_seq[

```r
ggplot(iver_data,
    aes(x = effectiveParties,
        y = povertyReduction,
        color = partySystem)) +
  geom_point(
  ) +
  geom_smooth(
    method = MASS::rlm
  )
```
]
 
.panel2-iver1-non_seq[
![](lab3_slides_files/figure-html/iver1_non_seq_01_output-1.png)
]

---
count: false
 
## Why are there multiple smooth curves? -- inherit.aes
.panel1-iver1-non_seq[

```r
ggplot(iver_data,
    aes(x = effectiveParties,
        y = povertyReduction,
        color = partySystem)) +
  geom_point(
*   inherit.aes = TRUE,
*   aes(x = effectiveParties,
*       y = povertyReduction,
*       color = partySystem)
  ) +
  geom_smooth(
    method = MASS::rlm
  )
```
]
 
.panel2-iver1-non_seq[
![](lab3_slides_files/figure-html/iver1_non_seq_02_output-1.png)
]

---
count: false
 
## Why are there multiple smooth curves? -- inherit.aes
.panel1-iver1-non_seq[

```r
ggplot(iver_data,
    aes(x = effectiveParties,
        y = povertyReduction,
        color = partySystem)) +
  geom_point(
    inherit.aes = TRUE,
    aes(x = effectiveParties,
        y = povertyReduction,
        color = partySystem)
  ) +
  geom_smooth(
*   inherit.aes = TRUE,
*   aes(x = effectiveParties,
*       y = povertyReduction,
*       color = partySystem),
    method = MASS::rlm
  )
```
]
 
.panel2-iver1-non_seq[
![](lab3_slides_files/figure-html/iver1_non_seq_03_output-1.png)
]

---

count: false
 
## You can either overwrite the geom_smooth() layer
.panel1-iver2-non_seq[

```r
ggplot(iver_data,
    aes(x = effectiveParties,
        y = povertyReduction,
        color = partySystem)) +
  geom_point(
  ) +
  geom_smooth(
    method = MASS::rlm
  )
```
]
 
.panel2-iver2-non_seq[
![](lab3_slides_files/figure-html/iver2_non_seq_01_output-1.png)
]

---
count: false
 
## You can either overwrite the geom_smooth() layer
.panel1-iver2-non_seq[

```r
ggplot(iver_data,
    aes(x = effectiveParties,
        y = povertyReduction,
        color = partySystem)) +
  geom_point(
  ) +
  geom_smooth(
*   aes(group = 1),
    method = MASS::rlm
  )
```
]
 
.panel2-iver2-non_seq[
![](lab3_slides_files/figure-html/iver2_non_seq_02_output-1.png)
]

---

count: false
 
## Or localize color aesthetics to the geom_point layer()
.panel1-iver3-non_seq[

```r
ggplot(iver_data,
    aes(x = effectiveParties,
        y = povertyReduction)) +
  geom_point(
  ) +
  geom_smooth(
    method = MASS::rlm
  )
```
]
 
.panel2-iver3-non_seq[
![](lab3_slides_files/figure-html/iver3_non_seq_01_output-1.png)
]

---
count: false
 
## Or localize color aesthetics to the geom_point layer()
.panel1-iver3-non_seq[

```r
ggplot(iver_data,
    aes(x = effectiveParties,
        y = povertyReduction)) +
  geom_point(
*   aes(color = partySystem)
  ) +
  geom_smooth(
    method = MASS::rlm
  )
```
]
 
.panel2-iver3-non_seq[
![](lab3_slides_files/figure-html/iver3_non_seq_02_output-1.png)
]

---

## A segue to a customized ggplot2 theme

- Chris and I wrote a customized ggplot2 theme to implement a minimalist aesthetics

- Can be downloaded here: https://students.washington.edu/kpleung/vis/theme_cavis.R

- To use it

```r
source("https://students.washington.edu/kpleung/vis/theme_cavis.R")
# source(".../theme_cavis.R") if you have download and put it in your directory
```

- I'll use this theme throughout my vis; you don't have to (and suggestions are welcomed!)

---

```r
ggplot(iris,
       aes(x = Sepal.Length,
           y = Petal.Length,
           color = Species)) +
  geom_point() +
* theme_grey()
```
]
 
.panel2-theme_cavis-rotate[
![](lab3_slides_files/figure-html/theme_cavis_rotate_01_output-1.png)
]

---
count: false
 
## Showcase customized theme
.panel1-theme_cavis-rotate[

```r
ggplot(iris,
       aes(x = Sepal.Length,
           y = Petal.Length,
           color = Species)) +
  geom_point() +
* theme_cavis_hgrid
```
]
 
.panel2-theme_cavis-rotate[
![](lab3_slides_files/figure-html/theme_cavis_rotate_02_output-1.png)
]

---
count: false
 
## Showcase customized theme
.panel1-theme_cavis-rotate[

```r
ggplot(iris,
       aes(x = Sepal.Length,
           y = Petal.Length,
           color = Species)) +
  geom_point() +
* theme_cavis_vgrid
```
]
 
.panel2-theme_cavis-rotate[
![](lab3_slides_files/figure-html/theme_cavis_rotate_03_output-1.png)
]

---
count: false
 
## Showcase customized theme
.panel1-theme_cavis-rotate[

```r
ggplot(iris,
       aes(x = Sepal.Length,
           y = Petal.Length,
           color = Species)) +
  geom_point() +
* theme_cavis
```
]
 
.panel2-theme_cavis-rotate[
![](lab3_slides_files/figure-html/theme_cavis_rotate_04_output-1.png)
]

---

## Visualization exercises: Electric vehicles in WA state

- Data are taken from [Washington State Department of Licensing](https://data.wa.gov/Transportation/Electric-Vehicle-Population-Data/f6w7-q2d2)

- "This dataset shows the Battery Electric Vehicles (BEVs) and Plug-in Hybrid Electric Vehicles (PHEVs) that are currently registered through Washington State Department of Licensing (DOL)."

- Load the electric vehicles data

```r
ev_data <- read_csv("data/ev_data.csv")
head(ev_data)
```

```
# A tibble: 6 × 8
  county    city      make  model model_year ev_type electric_range retail_price
  <chr>     <chr>     <chr> <chr>      <dbl> <chr>            <dbl>        <dbl>
1 Kittitas  Cle Elum  TESLA MODE…       2020 Batter…            291            0
2 Chelan    Chelan    TESLA MODE…       2020 Batter…            322            0
3 Snohomish Snohomish TESLA MODE…       2018 Batter…            215            0
4 Thurston  Tumwater  FORD  C-MAX       2014 Plug-i…             19            0
5 Kitsap    Kingston  FORD  FUSI…       2013 Plug-i…             19            0
6 King      Seattle   NISS… LEAF        2018 Batter…            151            0
```

---

## Visualization exercises: Electric vehicles in WA state

- Also some auxiliary county data with number of households and their median income

```r
county_data <- read_csv("data/county_data.csv")
head(county_data)
```

```
# A tibble: 6 × 3
  county      hh_num hh_income
  <chr>        <dbl>     <dbl>
1 King        789232     68065
2 San Juan      7613     50726
3 Snohomish   268325     66300
4 Kitsap       97220     59549
5 Washington 2620076     57244
6 Thurston    100650     60930
```

---

count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county1-auto[

```r
# Count EVs by county
*ev_data
```
]
 
.panel2-ev_by_county1-auto[

```
# A tibble: 114,312 × 8
   county    city     make  model model_year ev_type electric_range retail_price
   <chr>     <chr>    <chr> <chr>      <dbl> <chr>            <dbl>        <dbl>
 1 Kittitas  Cle Elum TESLA MODE…       2020 Batter…            291            0
 2 Chelan    Chelan   TESLA MODE…       2020 Batter…            322            0
 3 Snohomish Snohomi… TESLA MODE…       2018 Batter…            215            0
 4 Thurston  Tumwater FORD  C-MAX       2014 Plug-i…             19            0
 5 Kitsap    Kingston FORD  FUSI…       2013 Plug-i…             19            0
 6 King      Seattle  NISS… LEAF        2018 Batter…            151            0
 7 Skagit    Mount V… AUDI  E-TR…       2019 Batter…            204            0
 8 Kitsap    Port Or… NISS… LEAF        2019 Batter…            150            0
 9 Snohomish Bothell  TESLA MODE…       2020 Batter…            322            0
10 Yakima    Tieton   NISS… LEAF        2019 Batter…            150            0
# ℹ 114,302 more rows
```
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county1-auto[

```r
# Count EVs by county
ev_data %>%
* count(county)
```
]
 
.panel2-ev_by_county1-auto[

```
# A tibble: 39 × 2
   county       n
   <chr>    <int>
 1 Adams       35
 2 Asotin      46
 3 Benton    1394
 4 Chelan     654
 5 Clallam    741
 6 Clark     6827
 7 Columbia    13
 8 Cowlitz    576
 9 Douglas    227
10 Ferry       27
# ℹ 29 more rows
```
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county1-auto[

```r
# Count EVs by county
ev_data %>%
  count(county) %>%
* mutate(county = fct_reorder(county, n))
```
]
 
.panel2-ev_by_county1-auto[

```
# A tibble: 39 × 2
   county       n
   <fct>    <int>
 1 Adams       35
 2 Asotin      46
 3 Benton    1394
 4 Chelan     654
 5 Clallam    741
 6 Clark     6827
 7 Columbia    13
 8 Cowlitz    576
 9 Douglas    227
10 Ferry       27
# ℹ 29 more rows
```
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county1-auto[

```r
# Count EVs by county
ev_data %>%
  count(county) %>%
  mutate(county = fct_reorder(county, n)) %>%
* arrange(desc(county))
```
]
 
.panel2-ev_by_county1-auto[

```
# A tibble: 39 × 2
   county        n
   <fct>     <int>
 1 King      60059
 2 Snohomish 12720
 3 Pierce     8684
 4 Clark      6827
 5 Thurston   4207
 6 Kitsap     3902
 7 Whatcom    2873
 8 Spokane    2835
 9 Benton     1394
10 Island     1307
# ℹ 29 more rows
```
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county1-auto[

```r
# Count EVs by county
ev_data %>%
  count(county) %>%
  mutate(county = fct_reorder(county, n)) %>%
  arrange(desc(county)) ->
* ev_by_county
```
]
 
.panel2-ev_by_county1-auto[

]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county1-auto[

```r
# Count EVs by county
ev_data %>%
  count(county) %>%
  mutate(county = fct_reorder(county, n)) %>%
  arrange(desc(county)) ->
  ev_by_county

# Load scales package for labels and breaks
*library(scales)
```
]
 
.panel2-ev_by_county1-auto[

]

---

count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county2-non_seq[

```r
ggplot(ev_by_county,
       aes(y = county, x = n)) +
  theme_cavis_vgrid +
  geom_point()
```
]
 
.panel2-ev_by_county2-non_seq[
![](lab3_slides_files/figure-html/ev_by_county2_non_seq_01_output-1.png)
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county2-non_seq[

```r
ggplot(ev_by_county,
       aes(y = county, x = n)) +
  theme_cavis_vgrid +
  geom_point() +
* geom_col(
* )
```
]
 
.panel2-ev_by_county2-non_seq[
![](lab3_slides_files/figure-html/ev_by_county2_non_seq_02_output-1.png)
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county2-non_seq[

```r
ggplot(ev_by_county,
       aes(y = county, x = n)) +
  theme_cavis_vgrid +
  geom_point() +
  geom_col(
*   width = 0.05,
*   alpha = 0.5
  )
```
]
 
.panel2-ev_by_county2-non_seq[
![](lab3_slides_files/figure-html/ev_by_county2_non_seq_03_output-1.png)
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county2-non_seq[

```r
ggplot(ev_by_county,
       aes(y = county, x = n)) +
  theme_cavis_vgrid +
  geom_point() +
  geom_col(
    width = 0.05,
    alpha = 0.5
  ) +
* scale_x_continuous(
*   trans = "log10",
* )
```
]
 
.panel2-ev_by_county2-non_seq[
![](lab3_slides_files/figure-html/ev_by_county2_non_seq_04_output-1.png)
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county2-non_seq[

```r
ggplot(ev_by_county,
       aes(y = county, x = n)) +
  theme_cavis_vgrid +
  geom_point() +
  geom_col(
    width = 0.05,
    alpha = 0.5
  ) +
  scale_x_continuous(
    trans = "log10",
*   breaks = 10^(0:5),
*   labels = scales::label_comma(),
  )
```
]
 
.panel2-ev_by_county2-non_seq[
![](lab3_slides_files/figure-html/ev_by_county2_non_seq_05_output-1.png)
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county2-non_seq[

```r
ggplot(ev_by_county,
       aes(y = county, x = n)) +
  theme_cavis_vgrid +
  geom_point() +
  geom_col(
    width = 0.05,
    alpha = 0.5
  ) +
  scale_x_continuous(
    trans = "log10",
    breaks = 10^(0:5),
    labels = scales::label_comma(),
  ) +
* geom_text(
*   aes(label = n),
* )
```
]
 
.panel2-ev_by_county2-non_seq[
![](lab3_slides_files/figure-html/ev_by_county2_non_seq_06_output-1.png)
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county2-non_seq[

```r
ggplot(ev_by_county,
       aes(y = county, x = n)) +
  theme_cavis_vgrid +
  geom_point() +
  geom_col(
    width = 0.05,
    alpha = 0.5
  ) +
  scale_x_continuous(
    trans = "log10",
    breaks = 10^(0:5),
    labels = scales::label_comma(),
  ) +
  geom_text(
    aes(label = n),
*   nudge_x = 0.15
  )
```
]
 
.panel2-ev_by_county2-non_seq[
![](lab3_slides_files/figure-html/ev_by_county2_non_seq_07_output-1.png)
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county2-non_seq[

```r
ggplot(ev_by_county,
       aes(y = county, x = n)) +
  theme_cavis_vgrid +
  geom_point() +
  geom_col(
    width = 0.05,
    alpha = 0.5
  ) +
  scale_x_continuous(
    trans = "log10",
    breaks = 10^(0:5),
    labels = scales::label_comma(),
  ) +
  geom_text(
*   data = slice_max(ev_by_county, n, n = 10),
    aes(label = n),
    nudge_x = 0.15
  )
```
]
 
.panel2-ev_by_county2-non_seq[
![](lab3_slides_files/figure-html/ev_by_county2_non_seq_08_output-1.png)
]

---
count: false
 
## EV in WA state: Which county has most EVs?
.panel1-ev_by_county2-non_seq[

```r
ggplot(ev_by_county,
       aes(y = county, x = n)) +
  theme_cavis_vgrid +
  geom_point() +
  geom_col(
    width = 0.05,
    alpha = 0.5
  ) +
  scale_x_continuous(
    trans = "log10",
    breaks = 10^(0:5),
    labels = scales::label_comma(),
  ) +
  geom_text(
    data = slice_max(ev_by_county, n, n = 10),
    aes(label = n),
    nudge_x = 0.15
  ) +
* labs(y = NULL, x = "Total number of EVs")
```
]
 
.panel2-ev_by_county2-non_seq[
![](lab3_slides_files/figure-html/ev_by_county2_non_seq_09_output-1.png)
]

---

count: false
 
## EV in WA state: Trend of number of EVs across county
.panel1-ev_by_countyYear1-auto[

```r
# Count number of EVs by county
*ev_data
```
]
 
.panel2-ev_by_countyYear1-auto[

---
count: false
 
## EV in WA state: Trend of number of EVs across county
.panel1-ev_by_countyYear1-auto[

```r
# Count number of EVs by county
ev_data %>%
* group_by(model_year, county)
```
]
 
.panel2-ev_by_countyYear1-auto[

```
# A tibble: 114,312 × 8
# Groups:   model_year, county [497]
   county    city     make  model model_year ev_type electric_range retail_price
   <chr>     <chr>    <chr> <chr>      <dbl> <chr>            <dbl>        <dbl>
 1 Kittitas  Cle Elum TESLA MODE…       2020 Batter…            291            0
 2 Chelan    Chelan   TESLA MODE…       2020 Batter…            322            0
 3 Snohomish Snohomi… TESLA MODE…       2018 Batter…            215            0
 4 Thurston  Tumwater FORD  C-MAX       2014 Plug-i…             19            0
 5 Kitsap    Kingston FORD  FUSI…       2013 Plug-i…             19            0
 6 King      Seattle  NISS… LEAF        2018 Batter…            151            0
 7 Skagit    Mount V… AUDI  E-TR…       2019 Batter…            204            0
 8 Kitsap    Port Or… NISS… LEAF        2019 Batter…            150            0
 9 Snohomish Bothell  TESLA MODE…       2020 Batter…            322            0
10 Yakima    Tieton   NISS… LEAF        2019 Batter…            150            0
# ℹ 114,302 more rows
```
]

---
count: false
 
## EV in WA state: Trend of number of EVs across county
.panel1-ev_by_countyYear1-auto[

```r
# Count number of EVs by county
ev_data %>%
  group_by(model_year, county) %>%
* count()
```
]
 
.panel2-ev_by_countyYear1-auto[

```
# A tibble: 497 × 3
# Groups:   model_year, county [497]
   model_year county        n
        <dbl> <chr>     <int>
 1       1997 Snohomish     1
 2       1998 Clallam       1
 3       1999 Pierce        1
 4       1999 Skagit        1
 5       1999 Whatcom       1
 6       2000 King          4
 7       2000 San Juan      2
 8       2000 Snohomish     1
 9       2000 Whatcom       2
10       2002 Clallam       1
# ℹ 487 more rows
```
]

---
count: false
 
## EV in WA state: Trend of number of EVs across county
.panel1-ev_by_countyYear1-auto[

```r
# Count number of EVs by county
ev_data %>%
  group_by(model_year, county) %>%
  count() ->
* ev_by_countyYear
```
]
 
.panel2-ev_by_countyYear1-auto[

]

---

count: false
 
## EV in WA state: Trend of number of EVs across county
.panel1-ev_by_countyYear2-auto[

```r
*ggplot(ev_by_countyYear,
*      aes(x = model_year,
*          y = n,
*          group = county,
*          color = county == "King"))
```
]
 
.panel2-ev_by_countyYear2-auto[
![](lab3_slides_files/figure-html/ev_by_countyYear2_auto_01_output-1.png)
]