The Data

The data this week comes from Cleary et al, 2016 with the corresponding article Avian Assemblages at Bird Baths: A Comparison of Urban and Rural Bird Baths in Australia.

One note on the “cleaned” data is that the cleaning script didn’t account for the fact that there were totals in the Excel speadsheet (which should be removed before pivor_longer).

bird_baths <- readxl::read_xlsx("S1File.xlsx") %>%
  janitor::clean_names() %>%
  filter(!is.na(survey_year)) %>%
  mutate(survey_year = as.character(survey_year),
         observer = as.character(observer)) %>%
  mutate_if(is.numeric, as.logical)

Data Viz

With only 5 variables, there isn’t really much to model. So I’ll focus on a data visualization today.

I also wanted to try out a plot I saw recently here. It is a Venn diagram made into a bar chart! The R package is called ggupset.

The idea is, how many observers saw each of the bird types.

bird_baths_long <- bird_baths %>%
  pivot_longer(cols = 5:last_col(), names_to = "bird_type", values_to = "bird_count") 

bird_baths_long_true <- bird_baths_long %>%
  filter(bird_count) %>%
  select(-bird_count)
common_birds <- bird_baths_long_true %>%
  group_by(bird_type) %>%
  summarize(n_birds = n()) %>%
  filter(n_birds > 150) %>%
  pull(bird_type)
bird_baths_long_true %>%
  filter(bird_type %in% common_birds) %>%
  group_by(observer) %>%
  summarize(birds = list(bird_type)) %>%
  ggplot(aes(x = birds)) + 
  geom_bar() + 
  ggupset::scale_x_mergelist(sep = "-") +
  axis_combmatrix(sep = "-")
Bar chart with the height of the bar representing the number of observes who saw a particular type of bird. The bar chart is similar to a Venn diagram where the tallest bars represent the single birds sigtings (with or without others) and the lowest bars represent the number of sightings of a combination of many of the birds. The bars are ordered lexicographically.

For the 10 most common birds, how often were they seen together by a particular observer. 

bird_baths_long_true %>%
  filter(bird_type %in% common_birds) %>%
  group_by(observer) %>%
  summarize(birds = list(bird_type)) %>%
  ggplot(aes(x = birds)) + 
  geom_bar() + 
  scale_x_upset(order_by = "degree", n_intersections = 50)
Bar chart with the height of the bar representing the number of observes who saw a particular type of bird. The bar chart is similar to a Venn diagram where the tallest bars represent the single birds sigtings (with or without others) and the lowest bars represent the number of sightings of a combination of many of the birds. The bars are ordered by degree.

For the 10 most common birds, how often were they seen together by a particular observer. 

praise()
## [1] "You are unreal!"