Skip to content
This repository was archived by the owner on May 10, 2022. It is now read-only.

Latest commit

 

History

History
152 lines (116 loc) · 5.17 KB

transition_manual.md

File metadata and controls

152 lines (116 loc) · 5.17 KB

transition_manual

Adam Gruer + Saskia Freytag 22/11/2018

Explain transition_manual

A static plot

cars_plot <- ggplot(mtcars, aes(disp, mpg), ) + 
  geom_point(colour = "purple", size = 3) 
cars_plot

Facet by cylinder

cars_plot + facet_wrap(~cyl)

##Animate

Transition manual will show one frame per level of the supplied variable. The {current_frame} can be used anywhere that accepts a string (I think!) to display the value of the ‘frame’ variable at each step of the animation.

cars_plot + transition_manual(cyl) +
  labs(title = "{current_frame}")
## nframes and fps adjusted to match transition

More complex example

In this example we will be using the okcupiddata 📦 ❤️ to visualize how feelings about pets change with star sign, sex and relationship status. First let’s begin with some data wrangling.

Intitally, we extract the columns we are interested in in the profiles data.

data(profiles)
profiles_red <- profiles %>% select(pets, sex, status, sign)

Unfortunatley, the sign column does not simply contain the 12 star signs, but includes people’s feelings about the importance of star signs. We will delete these for now. We will then ensure that the sign column is a factor with ordered levels for the different star signs.

profiles_red <- profiles_red %>% mutate(sign = str_extract(sign, "[a-z']+[[:space:]]")) %>% mutate(sign = gsub(" ", "", sign))
profiles_red <- profiles_red %>% 
  mutate(sign = factor(sign, levels=c("aries", "taurus", "gemini", "cancer",
                                      "leo", "virgo", "libra", "scorpio",
                                      "sagittarius", "capricorn", "aquarius",
                                      "pisces")))

The next step is to entangle the pets column, which includes people’s likes and dislikes of pets as well as their pet ownership. We will disentangle all of these into three new columns that we will call pets_dislike, pets_like and pets_has.

profiles_red <- profiles_red %>% 
  mutate(pets_dislike = case_when(
    str_detect(pets, "dislikes dogs and dislikes cats") ~ "dogs_cats",
    str_detect(pets, "dislikes dogs") ~ "dogs", 
    str_detect(pets, "dislikes cats") ~ "cats"))
profiles_red <- profiles_red %>% mutate(sign = gsub(" ", "", sign))
profiles_red <- profiles_red %>% 
  mutate(pets_like = case_when(
    str_detect(pets, "likes dogs and likes cats") ~ "dogs_cats",
    str_detect(pets, "likes dogs") ~ "dogs", 
    str_detect(pets, "likes cats") ~ "cats"))
profiles_red <- profiles_red %>% 
  mutate(pets_has = case_when(
    str_detect(pets, "has dogs and has cats") ~ "dogs_cats",
    str_detect(pets, "has dogs") ~ "dogs", 
    str_detect(pets, "has cats") ~ "cats"))

Also let’s get rid of individuals that did not answer the star sign question.

profiles_red <- profiles_red %>% filter(!is.na(sign))

We will visualize individuals that dislike pets first. In order to make it look way, way cooler, we will use the ggtextures 📦 to add some much needed fun into this plot. This is super easy as all you have to do is to find some images you like and use the wonderful functions geom_textured_bar in combination with scale_image_manual.

profiles_red2 <- profiles_red %>% filter(!is.na(pets_dislike))

images = c(
 dogs= "https://www.doggroomingtutorial.com/wp-content/uploads/2013/03/hairless-dogs.jpg",
 cats= "https://cf.ltkcdn.net/cats/images/std/143224-306x425-Full-shot-of-Sphynx-cat.jpg",
 dogs_cats = "https://us.123rf.com/450wm/fotojagodka/fotojagodka1704/fotojagodka170400004/76409120-canadian-sphynx-cat-and-thai-ridgeback-dog-in-front-of-white-background.jpg?ver=6"
)

dislike_plot <- ggplot(profiles_red2, aes(sex, image = pets_dislike)) +
  geom_textured_bar() +
  scale_image_manual(values = images)

dislike_plot + transition_manual(sign) + labs(title="{current_frame}")
## nframes and fps adjusted to match transition

Finally, we will visualize peoples like of pets by their relationship status and star sign. Instead of doing it with counts, we will visualize the proportions!

profiles_red3 <- profiles_red %>% filter(!is.na(pets_like))
like_plot <- ggplot(profiles_red3, aes(x = status)) + geom_bar(aes(fill=pets_like), position="fill") 

images = c(
 dogs= "https://www.akc.org/wp-content/themes/akc/component-library/assets//img/welcome.jpg",
 cats= "https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcS2WhOti17crdhQ5YaH0tWnaPwy-EL7rVEWpyH_VABFAmjfA8mZKA",
 dogs_cats = "https://www.southernmallee.sa.gov.au/webdata/resources/images/Cat_and_Dog.jpg"
)

like_plot <- ggplot(profiles_red3, aes(status, image = pets_like, class=pets_like)) +
  geom_textured_bar(position="fill") +
  scale_image_manual(values = images) +
  ylab("proportion")

like_plot + transition_manual(sign) + labs(title="{current_frame}")
## nframes and fps adjusted to match transition