36-315-Lec8

class: inverse, center, middle

# 36-315: Statistical Graphics and Visualization
## Lecture 8

Meghan Hall Department of Statistics & Data Science Carnegie Mellon University June 9, 2021

---
layout: true
<div class="my-footer">cmu-36315.netlify.app</div>
---

# From last time

.large[Grab bag!]
 
.medium[maps, pie charts, heat maps]

.large[Data manipulation]
 
.medium[aggregating, joining, pivoting]

---

# Today

.large[Taking plots to the next level]
 
.medium[effective *and* elegant]

.large[Redoing plots we've seen before]
 
.medium[get familiar with the syntax and the options]

---

# Intro

.large[Previously: exploratory data viz]
 
.medium[aesthetics are less important, as long as you're not distorting your data]
 
 
--
 
.large[Moving toward explanatory data viz for presentation]
 
.medium[data viz = communication]
 
 
--
 
.large[What's necessary vs. what's aesthetic preference]
 
.medium[there are always bad options, but there's rarely only one good option]

---

# Data-ink ratio

.large[The ratio of your data to everything else]
 
.medium[popularized by Edward Tufte]
 
 
--
 
.large[Be mindful, but don't go too far]
 
 
--
 
.large[The "extras" may add visual clutter, technically, but they're essential for orienting yourself]
 
.medium[gridlines, axes, etc.]

---

# Titles

.large[The purpose is to make a point!]
 
 
--
 
.large[Don't be afraid to put your conclusion in your title]
 
.medium[guide your reader to the purpose of the graph]
 
.medium[why am I making this plot?]
 
 
--
 
.large[Use subtitles and captions to add extra context or data source info]

---

# Axes

.large[When axis labels aren't necessary]
 
.medium[generally: time, regions, many categorical variables]
 
 
--
 
.large[Adjust font size!]
 
.medium[frequently knit to your desired output to check scale]
 
.medium[axis labels are often too small]

---

# Legends

.large[Limit their presence when possible]
 
.medium[move onto the plot if there's space]
 
.medium[label relevant points instead]
 
.medium[or match colors to words in the title]
 
 
--
 
.large[Order in a meaningful way]
 
.medium[match ordering of lines to the legend]

---

# Multiple plots

.large[Maintain the same color scheme throughout]
 
.medium[reuse/share legends where possible]
 
 
--
 
.large[Move from simple to more complex]
 
 
--
 
.large[Diversity chart types]
 
 
--
 
.large[Generally one "point" per graph]

---

# Syntax

.large[Never worry about memorizing all the syntax]
 
.medium[Google is your friend]
 
 
--
 
.large[Be aware of the guidelines we'll discuss]
 
.medium[and how out-of-the-box ggplots can fall short]
 
 
--
 
.large[Know what options are available so you can look up details]

---

# Some golden rules of graphs

Don’t add complexity without a good reason.

Everything (everything!) must be readable.

Don’t distort data, intentionally or not.

Be mindful of the data-to-ink ratio.

All axes, labels, etc. should have real titles, not code variable names.

Always strive for clarity.

Titles, subtitles, and captions should add information.

---

# 1: HW 1, Problem 2

---

# 1: HW 1, Problem 2

---

.h-1[# 1: HW 1, Problem 2]

.tiny[

```r
mpg %>% 
  filter(year == 2008 & class == "suv") %>% 
  group_by(manufacturer) %>% 
  summarize(median = median(cty),
            n = n()) %>% 
* mutate(manufacturer = str_to_title(manufacturer),
*        label = paste0(manufacturer, "\n", n,
*                       ifelse(n > 1, " models", " model"))) %>%
  ggplot(aes(x = reorder(label, -median), y = median)) +
  geom_bar(stat = "identity") +
* scale_y_continuous(expand = expansion(mult = c(0, .1))) +
* labs(title = "Subaru SUVs have highest median city mpg",
       subtitle = "Among manufacturers in 2008",
       x = NULL,
       y = "Miles per gallon") +
* theme(axis.ticks = element_blank(),
*       panel.background = element_blank(),
*       panel.grid.major.y = element_line(color = "grey90", size = 0.2),
*       panel.border = element_rect(color = "black", fill = NA, size = 0.5))
```
]

---

# 2: HW 1, Problem 3

---

# 2: HW 1, Problem 3

---

.h-1[# 2: HW 1, Problem 3]

.tiny[

```r
mpg %>% 
  filter(manufacturer %in% c("toyota","dodge","audi","nissan") & year == 2008) %>% 
  group_by(manufacturer, drv) %>% 
  summarize(mean = mean(hwy)) %>% 
* mutate(manufacturer = str_to_title(manufacturer),
*        drv = fct_recode(drv,  "4-wheel drive" = "4", "front-wheel drive" = "f")) %>%
  ggplot(aes(x = manufacturer, y = mean, fill = drv)) +
  geom_bar(stat = "identity", position = "dodge") +
* scale_fill_manual(values = c("#707078", "#BB0000"), name = NULL) +
  geom_text(aes(label = round(mean, 1)), position = position_dodge(width = 0.9), 
*           vjust = 1.5, color = "white", fontface = 2) +
* labs(title = "Average highway miles per gallon is higher in front-wheel drive vehicles",
       subtitle = "Among selected manufacturers in 2008",
       x = NULL,
       y = "Miles per gallon") +
  theme(axis.ticks = element_blank(),
        panel.background = element_blank(),
        panel.grid.major.y = element_line(color = "grey90", size = 0.2),
        panel.border = element_rect(color = "black", fill = NA, size = 0.5),
*       legend.position = c(0.44, 0.85),
*       legend.text = element_text(size = 11))
```
]

---

# 3: Lab 3, Problem 3

---

# 3: Lab 3, Problem 3

---

.h-1[# 3: Lab 3, Problem 3]

.tiny[

```r
txhousing %>% 
  filter(year >= 2010) %>% 
  ggplot(aes(x = as.character(year), y = median)) +
  geom_violin(draw_quantiles = c(0.25, 0.5, 0.75)) +
* geom_jitter(alpha = .25, width = .3, size = 0.5, color = "#bb0000") +
* scale_y_continuous(labels = dollar, breaks = seq(100000, 300000, 50000)) +
  labs(title = "The distribution of median home prices by city in Texas") +
  theme(axis.title = element_blank(),
        panel.background = element_blank(),
        panel.grid.major.y = element_line(color = "grey90", size = 0.2),
        panel.border = element_rect(color = "black", fill = NA, size = 0.5),
        axis.ticks = element_blank(),
*       axis.text = element_text(size = 10, face = 2),
*       plot.title.position = "plot")
```
]

---

# 4: Lecture 5

---

# 4: Lecture 5

---

.h-1[# 4: Lecture 5]

.tiny[

```r
constructor_pts %>% 
  filter(year == 2020) %>% 
  mutate(third = ifelse(name %in% 
                          c("McLaren","Renault","Racing Point"), name, "z"),
         label = if_else(round == max(round) &
                           third != "z", name, NA_character_)) %>%
  ggplot(aes(x = round, y = points, color = third, group = name, 
             size = third, alpha = third)) + 
  geom_line() +
  geom_label_repel(aes(label = label), size = 4.5) +
  scale_x_continuous(breaks = seq(1, 17, 1)) +
  scale_color_manual(values = c("#E0610E","#F596C8",
                                "#FFF500","dark grey")) +
  scale_size_manual(values = c(1.5, 1.5, 1.5, 0.75)) +
  scale_alpha_manual(values = c(1, 1, 1, 0.3)) +
* labs(title = "The race for third place during the 2020 F1 season",
*      subtitle = "While Mercedes and Red Bull ran off with the first two placings, three teams battled all year long for third place",
       x = "Race Round",
       y = "Accumulated Points") +
  theme(legend.position = "none",
        panel.background = element_blank(),
*       panel.grid.major.y = element_line(color = "grey90", size = 0.2),
        axis.ticks = element_blank(),
        panel.border = element_rect(color = "black", fill = NA, size = 0.5))
```
]

---

# 5: Lecture 6

---

# 5: Lecture 6

---

.h-1[# 5: Lecture 6]

.tiny[

```r
friends_info %>% 
* mutate(label = ifelse(us_views_millions > 50, title, NA_character_)) %>%
  ggplot(aes(x = us_views_millions, y = imdb_rating, color = season)) + 
  geom_jitter(size = 2) +
  scale_colour_gradient(low = "#fafafa",high = "#191970",breaks = seq(1, 10, 1),
                        name = "Season") +
* geom_label_repel(aes(label = label, x = us_views_millions,
*                      y = imdb_rating), size = 4, inherit.aes = FALSE) +
* scale_x_continuous(labels = label_number(suffix = "M")) +
  scale_y_continuous(breaks = seq(7, 10, 0.5)) +
* labs(title = "Two Friends episodes were viewed far more than the rest",
*      subtitle = "Both highlighted episodes were two-part episodes",
       x = "US views",
       y = "IMDB rating") +
* theme(legend.position = c(0.75, 0.08),
*       legend.direction = "horizontal",
*       legend.background = element_blank(),
*       legend.title = element_text(color = "#353839", size = 11, face = "bold", vjust = 0.75),
        axis.ticks = element_blank(),
        panel.background = element_blank(),
        panel.border = element_rect(color = "black", fill = NA, size = 0.5),
        panel.grid.major = element_line(color = "grey90", size = 0.3))
```
]

---

# 6: Lecture 6

---

# 6: Lecture 6

---

.h-1[# 6: Lecture 6]

.tiny[

```r
penguins %>% 
  filter(species != "Gentoo") %>% 
  mutate(label = case_when(flipper_length_mm == 192 & 
                             body_mass_g == 2700 ~ "Chinstrap",
                           flipper_length_mm == 184 & 
                             body_mass_g == 4650 ~ "Adelie")) %>% 
  ggplot(aes(x = flipper_length_mm, y = body_mass_g, 
             size = bill_length_mm, color = species)) +
  geom_point(alpha = 0.5) +
  scale_size(range = c(0.1, 7), breaks = c(35, 40, 45, 50, 55), 
             name = "Bill Length (mm)") +
* geom_label_repel(aes(x = flipper_length_mm, y = body_mass_g,
*                      color = species, label = label),
                   inherit.aes = FALSE) +
* scale_color_discrete(guide = "none") +
  labs(x = "Flipper Length (mm)",
       y = "Body Mass (g)",
       title = "Chinstrap penguins tend to have longer flippers and longer bills",
*      caption = "Data from the palmerpenguins package") +
  theme(legend.position = "top",
        panel.background = element_blank(),
*       panel.grid.major = element_line(color = "grey90", size = 0.2),
*       axis.ticks = element_line(color = "grey90", size = 0.2),
*       legend.key = element_rect(fill = "transparent"))
```
]

---

# 7: Lecture 7

---

# 7: Lecture 7

---

.h-1[# 7: Lecture 7]

.tiny[

```r
lincoln_weather %>% 
 select(CST, temp = `Max Temperature [F]`) %>% 
 mutate(date = ymd(CST), 
* month = month(date, label = TRUE),
 day = day(date)) %>% 
 ggplot(aes(x = month, y = day, fill = temp)) +
 geom_tile(color = "white") +
 scale_y_continuous(trans = "reverse", 
 breaks = seq(1, 31, 5)) + 
 labs(title = "Maximum Temperature by day in Lincoln, NE in 2016") +
* geom_text(aes(color = temp < 40, label = temp), size = 2) +
* scale_color_manual(guide = FALSE, values = c("black", "white")) +
 scale_fill_gradient(low = "blue", high = "yellow",
 breaks = seq(20, 90, 10),
 name = " °F") +
* guides(fill = guide_colorsteps()) +
 theme(axis.ticks = element_blank(),
* panel.background = element_blank(),
 axis.title = element_blank(),
* axis.text = element_text(face = 2)) +
 coord_cartesian(expand = FALSE)
```
]

---

# 8: Lecture 7

---

# 8: Lecture 7

---

.h-1[# 8: Lecture 7]

.tiny[

```r
penguins %>% 
  ggplot(aes(x = body_mass_g, y = ..count..)) +
  geom_density_line(data = select(penguins, -species), 
*                   aes(fill = "all penguins"), color = "transparent") +
* geom_density_line(aes(fill = "species"), color = "transparent") +
  facet_wrap(~species, nrow = 1) +
  scale_fill_manual(values = c("grey","#0C8346"), name = NULL,
                    guide = guide_legend(direction = "horizontal")) +
  labs(x = "Body Mass (g)",
*      title = "Comparing the distribution of body mass by penguin species",
*      subtitle = "Gentoo penguins tend to be the heaviest") +
  theme(legend.position = "bottom",
        panel.background = element_blank(),
*       panel.grid.major = element_line(color = "grey90", size = 0.3),
*       strip.text = element_text(face = "bold", color = "white"),
*       strip.background = element_rect(fill = "#0C8346"),
        panel.border = element_rect(color = "black", fill = NA, size = 0.5),
        axis.ticks = element_blank(),
*       plot.title.position = "plot")
```
]

---

# 9: HW 2, Problem 5

---

# 9: HW 2, Problem 5

---

.h-1[# 9: HW 2, Problem 5]

.tiny[

```r
penguins %>% 
  ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
  geom_point() +
* geom_smooth(method = "lm", show.legend = FALSE, alpha = 0.3) +
* labs(title = "Bill depth is positively correlated with bill length, regardless of species",
       y = "Bill Depth (mm)",
       x = "Bill Length (mm)") +
* guides(color = guide_legend(override.aes = list(size = 5))) +
  theme(legend.position = c(0.1, 0.1),
        legend.title = element_blank(),
        panel.background = element_blank(),
        panel.grid.major = element_line(color = "grey90", size = 0.3),
        axis.ticks = element_line(color = "grey90", size = 0.3),
*       legend.key = element_rect(fill = "transparent"))
```
]

---

# 10: Lab 5, Problem 1

---

# 10: Lab 5, Problem 1

---

.h-1[# 10: Lab 5, HW 1]

.tiny[

```r
msleep %>% 
 filter(vore %in% c("carni","herbi")) %>% 
 mutate(name = fct_reorder(name, sleep_total),
 name = fct_reorder(name, vore),
* vore = fct_recode(vore, "herbivore" = "herbi", "carnivore" = "carni")) %>%
 group_by(vore) %>% 
 mutate(mean = mean(sleep_total)) %>% 
 ggplot(aes(x = sleep_total, y = name, color = vore)) +
 geom_point(size = 2) +
 scale_color_manual(values = c("#bb0000","#098641")) +
* scale_x_continuous(name = NULL, labels = number_format(suffix = " hrs", accuracy = 1)) +
 labs(y = NULL, x = "Total sleep per day (hrs)",
* title = "**Carnivores** sleep slightly more per day than **herbivores**",
 subtitle = "Among these species, carnivores sleep one hour more per day on average") +
* theme(plot.title = element_markdown(),
* plot.subtitle = element_text(margin = margin(-5, 0, 10, 0)),
 plot.title.position = "plot",
 legend.position = "none",
 panel.background = element_blank(),
 plot.background = element_rect(fill = "transparent", color = NA),
 panel.grid.major = element_line(color = "black", size = 0.1),
 axis.ticks = element_blank(),
* axis.text.y = element_text(size = 6.75))
```
]

---

# Upcoming

.large[Lab 6 on Thursday June 10]

.large[Lecture 9 on Friday June 11]

.large[Homework 2 due Tuesday June 15]