1 Outline

Import Apple Mobility Data
Facets
Complete themes
Adjusting colors
geofacet

2 Materials

View the slides for this section here.

3 Import `TidyApple` data

We’re going to import the data from the previous exercises.

TidyApple <- vroom::vroom("https://bit.ly/3IiwcJs", delim = ",")
TopUSCities <- vroom::vroom("https://bit.ly/3d1zEd0", delim = ",")

4 Advanced Facets

In the previous lesson, we introduced the facet_wrap() function for viewing the relationship between two variables across the levels of a categorical variable. In the next section, we’re going to show how faceting can be used to explore ‘small multiples’ in a dataset with variation across multiple levels.

4.1 `facet_wrap()`

Now that we have a graph we can use to compare the 8 cities, we will use facet_wrap to create a subplot for each level of region.

4.1.1 exercise

Fill in the facet_wrap() (note the use of the ~) function with region and set the ncol to 2.

TopUSCities %>%
  ggplot(aes(x = date, y = dir_request,
             group = trans_type,
             color = trans_type)) +
  geom_line(size = 0.1) +
  geom_hline(yintercept = 100, size = 0.2, color = "gray20") +
  facet_wrap(~ _______,  ncol = _) +
  lab_top_cities

4.1.2 solution

lab_top_cities <- labs(x = "Date", y  = "Direction Requests",
     title = "Trends of Relative Activity in Selected US Cities",
     subtitle = "NY, LA, CH, HOU, PHA, PHL, SATX, SD",
     color = "Type")

TopUSCities %>% 
  ggplot(aes(x = date, y = dir_request,
             group = trans_type,
             color = trans_type)) +
  geom_line(size = 0.1) +
  geom_hline(yintercept = 100, size = 0.2, color = "gray20") +
  facet_wrap(~ region,  ncol = 2) +
  lab_top_cities

4.1.3 exercise

Now map both region and trans_type to facet_wrap() and set the ncol to 6.

TopUSCities %>%
  ggplot(aes(x = date, y = dir_request,
             group = region,
             color = trans_type)) +
  geom_line(size = 0.1) +
  geom_hline(yintercept = 100, size = 0.2, color = "gray20") +
  facet_wrap(_______ ~ _______, ncol = _) +
  lab_top_cities

4.1.4 solution

TopUSCities %>%
  ggplot(aes(x = date, y = dir_request,
             group = region,
             color = trans_type)) +
  geom_line(size = 0.1) +
  geom_hline(yintercept = 100, size = 0.2, color = "gray20") +
  facet_wrap(region ~ trans_type, ncol = 6) +
  lab_top_cities

4.2 Complete Themes

ggplot2 comes with a few default theme settings. Read more about them here. These themes come with arguments to adjust the font_size and font_family on our plot.

theme_minimal(
  base_size = 11,
  base_family = "",
  base_line_size = base_size/22,
  base_rect_size = base_size/22
)

4.2.1 exercise

We’ll use the theme_minimal() function to reduce the number of elements on our graph (don’t add any arguments).

TopUSCities %>%
  ggplot(aes(x = date, y = dir_request,
             group = region,
             color = trans_type)) +
  geom_line(size = 0.1) +
  geom_hline(yintercept = 100, size = 0.2, color = "gray20") +
  facet_wrap(region ~ trans_type, ncol = 6) +
  ______________() +
  lab_top_cities

4.2.2 solution

TopUSCities %>%
  ggplot(aes(x = date, y = dir_request,
             group = region,
             color = trans_type)) +
  geom_line(size = 0.1) +
  geom_hline(yintercept = 100, size = 0.2, color = "gray20") +
  facet_wrap(region ~ trans_type, ncol = 6) +
  theme_minimal() +
  lab_top_cities

4.3 Theme elements

Themes (accessible with the theme function) give up the ability to customize various element_() settings, which affect the ‘look’ of our graph.

4.3.1 exercise

We will start by moving the legend.position to "top" so it’s not crowding the x axis on our graph.

TopUSCities %>%
  ggplot(aes(x = date, y = dir_request,
             group = region,
             color = trans_type)) +
  geom_line(size = 0.1) +
  geom_hline(yintercept = 100, size = 0.2, color = "gray20") +
  facet_wrap(region ~ trans_type, ncol = 6) +
  theme_minimal() +
  theme(___________ = _____) +
  lab_top_cities

4.3.2 solution

TopUSCities %>%
  ggplot(aes(x = date, y = dir_request,
             group = region,
             color = trans_type)) +
  geom_line(size = 0.1) +
  geom_hline(yintercept = 100, size = 0.2, color = "gray20") +
  facet_wrap(region ~ trans_type, ncol = 6) +
  theme_minimal() +
  theme(legend.position = "top") +
  lab_top_cities

Notice we pass the theme() layer after the theme_minimal() layer (if we ordered these the other way around, theme_minimal() would overwrite any custom settings we made in theme()).

4.4 Small Multiples

What if we wanted to facet more than 8 cities? Fortunately, we have a the ggforce and geofacet packages for doing just that!

library(sf)
library(geofacet)
library(ggforce)
library(jcolors)

Building the Graph Data

We will start by filtering the TidyApple to only the 50 US states (we’ve removed three US territories) and storing these data in TidyAppleUS.
Next we limit the date range to the beginning of the shelter in place (from February 1, 2020 to May 1, 2020). These data get stored in TidyAppleUST1.
We then create a dataset with only "transit" direction requests, and we count these by state (sub_region), arrange the data descending with sort = TRUE, and take the top 25 rows (Top25TransitStates).

Finally, we filter TidyAppleUST1 using the 25 states in Top25TransitStates to create our graphing dataset, Top25TransitUSAllT1.

# create only US states (TidyAppleUS)
TidyApple %>%
  filter(country == "United States" &
         !sub_region %in% c("Guam", "Puerto Rico",
                            "Virgin Islands")) -> TidyAppleUS
# create shelter in place time 1 (TidyAppleUST1)
TidyAppleUS %>%
  filter(date >= as_date("2020-02-01") &
           date <= as_date("2020-05-01")) -> TidyAppleUST1

# create top 25 states (Top25TransitStates)
Top25TransitStates <- TidyAppleUST1 %>%
  filter(trans_type == "transit") %>%
  count(sub_region, trans_type, sort = TRUE) %>%
  head(25)

# filter T1 to states with the most transit requests (Top25TransitUSAllT1)
TidyAppleUST1 %>%
  filter(sub_region %in%
           unique(Top25TransitStates$sub_region)) -> Top25TransitUSAllT1

Top25TransitUSAllT1 %>% skimr::skim()

Data summary
Name	Piped data
Number of rows	188643
Number of columns	7
_______________________
Column type frequency:
character	5
Date	1
numeric	1
________________________
Group variables	None

Variable type: character

skim_variable	complete_rate	min	max	n_unique
geo_type	1	4	6	2
region	1	5	39	1058
trans_type	1	7	7	3
sub_region	1	4	14	25
country	1	13	13	1

Variable type: Date

skim_variable	n_missing	complete_rate	min	max	median	n_unique
date	0	1	2020-02-01	2020-05-01	2020-03-17	91

Variable type: numeric

skim_variable	n_missing	complete_rate	mean	sd	p0	p25	p50	p75	p100	hist
dir_request	0	1	92.68	36.02	0.44	68.64	93.6	113.14	1379.02	▇▁▁▁▁

4.4.1 exercise

set title to "States With Highest Transit Direction Requests"
set subtitle to "Top 25 states based on number of total transit requests"

lab_facet_wrap_paginate <- labs(
       x = "Date", y = "Direction Requests",
       title = _____________________________________,
       subtitle = _____________________________________)

4.4.2 solution

lab_facet_wrap_paginate <- labs(
       x = "Date", y = "Direction Requests",
       title = "States With Highest Transit Direction Requests",
       subtitle = "Top 25 states based on number of total transit requests")

4.4.3 exercise

Inside ggforce::facet_wrap_paginate():

map sub_region as the variable to facet using the ~
map 5 to ncol
map "free_y" to scales

Inside theme()

map element_blank() to panel.border and panel.background
map element_text(size = 6) to axis.text.x and axis.text.y
map element_text(colour = 'black') to strip.text
map element_rect(fill = "gray93") to strip.background
map "top" to legend.position

Top25TransitUSAllT1 %>%
  # global settings
  ggplot(aes(x = date, y = dir_request,
             group = trans_type,
             color = trans_type)) +
  # lines
  geom_hline(yintercept = 100, size = 0.3, color = "black") +
  geom_line(size = 0.2) +
  # faceting
  ggforce::facet_wrap_paginate(~ __________,
                               ncol = _,
                               scales = _______) +
  # theme settings
  theme(__________ = __________(),
        __________ = __________(),

        __________ = __________(size = _),
        __________ = __________(size = _),

        __________ = __________(colour = __________),

        __________ = __________(fill = __________),

        __________ = __________) +
  # labels
  lab_facet_wrap_paginate

4.4.4 solution

Top25TransitUSAllT1 %>%
  # global settings
  ggplot(aes(x = date, y = dir_request,
             group = trans_type,
             color = trans_type)) +
  # lines
  geom_hline(yintercept = 100, size = 0.3, color = "black") +
  geom_line(size = 0.2) +
  # faceting
  ggforce::facet_wrap_paginate(~ sub_region,
                               ncol = 5,
                               scales = "free_y") +
  # theme settings
  theme(panel.border = element_blank(),
        panel.background = element_blank(),
        axis.text.x = element_text(size = 6),
        axis.text.y = element_text(size = 6),
        strip.text = element_text(colour = 'black'),
        strip.background = element_rect(fill = "gray93"),
        legend.position = "top") +
  # labels
  lab_facet_wrap_paginate

4.5 Adjusting Colors

Changing the colors on graphs gives us the ability to further customize their look. We can set these manually, or use one of the many complete color palettes from a user-written package. Below we’ll use the jcolors package to highlight the transit direction requests from the previous graph.

4.5.1 exercise

add scale_color_jcolors() and play with the palette argument to make the graph look like the solution.

Top25TransitUSAllT1 %>%
  # global settings
  ggplot(aes(x = date, y = dir_request,
             group = trans_type,
             color = trans_type)) +
  # lines
  geom_hline(yintercept = 100, size = 0.3, color = "black") +
  geom_line(size = 0.2) +
  # faceting
  ggforce::facet_wrap_paginate(~ sub_region,
                               ncol = 5,
                               scales = "free_y") +
  # theme settings
  theme(panel.border = element_blank(),
        panel.background = element_blank(),
        axis.text.x = element_text(size = 6),
        axis.text.y = element_text(size = 6),
        strip.text = element_text(colour = 'black'),
        strip.background = element_rect(fill = "gray93"),
        legend.position = "top") +
  # adjust colors
  __________________________(palette = ____) +

  lab_facet_wrap_paginate

4.5.2 solution

Top25TransitUSAllT1 %>%
  # global settings
  ggplot(aes(x = date, y = dir_request,
             group = trans_type,
             color = trans_type)) +
  # lines
  geom_hline(yintercept = 100, size = 0.3, color = "black") +
  geom_line(size = 0.2) +
  # faceting
  ggforce::facet_wrap_paginate(~ sub_region,
                               ncol = 5,
                               scales = "free_y") +
  # theme settings
  theme(panel.border = element_blank(),
        panel.background = element_blank(),
        axis.text.x = element_text(size = 6),
        axis.text.y = element_text(size = 6),
        strip.text = element_text(colour = 'black'),
        strip.background = element_rect(fill = "gray93"),
        legend.position = "top") +

  scale_color_jcolors(palette = "pal3") +

  lab_facet_wrap_paginate

4.6 Extra: `geofacet`

We’re now going to look at all 50 states using the facet_geo() function from the geofacet package. To make this graph easier to interpret, we’re going to focus only on walking mobility data, and adjust the dir_request value to absolute change from baseline (set to 100 on 2020-01-13).

4.6.1 Adjusted Walking Data

Create the USWalkingAdj data by filtering the trans_type to "walking" and creating two new variables: above_below (a logical indicator for values being above or below the baseline value of 100), and dir_request_adj (the adjusted direction request value).

4.6.2 exercise

Fill in the correct variables in the wrangling steps below:

USWalkingAdj <- TidyAppleUS %>%
  filter(trans_type == _________) %>%
  mutate(above_below = _________ < 100,
         dir_request_adj = _________ - 100)

USWalkingAdj %>%
  skimr::skim()

4.6.3 solution

USWalkingAdj <- TidyAppleUS %>%
  filter(trans_type == "walking") %>%
  mutate(above_below = dir_request < 100,
         dir_request_adj = dir_request - 100)

USWalkingAdj %>%
  skimr::skim()

Data summary
Name	Piped data
Number of rows	160085
Number of columns	9
_______________________
Column type frequency:
character	5
Date	1
logical	1
numeric	2
________________________
Group variables	None