I came across this dataset from NYC from Urban Park Ranger Animal Condition Response datasource. From May 2018 - June 2019, data was collected on animal calls received by the Urban Park Rangers of New York City. Most of these reports would be associated with city parks for animals in distress.
Let’s take a look at the data and see what types of animals have appeared in NYC!
data_raw <- read.csv("Urban_Park_Ranger_Animal_Condition_Response.csv")
#Animal by species
head(data_raw %>% count(Species.Description, sort =T))
## Species.Description n
## 1 Raccoon 497
## 2 raccoon 64
## 3 Canada Goose 37
## 4 Chicken 26
## 5 Red-tailed Hawk 25
## 6 Opossum 17
#Animal by classification
head(data_raw %>% count(Animal.Class, sort =T))
## Animal.Class n
## 1 Small Mammals-RVS 582
## 2 Birds 166
## 3 Domestic 57
## 4 Raptors 54
## 5 Small Mammals-non RVS 49
## 6 Terrestrial Reptile or Amphibian 21
Clearly, a lot of calls for raccoons! Unfortunately, this dataset does not include the reason for the call besides healthy, unhealthy, dead on arrival (DOA). There is some cleaning needed when looking at the classifications. Looking at $Marine Reptiles shows they’re actually $Terrestrial Reptiles. Let’s do some cleaning.
data_clean <- data_raw %>%
mutate(
Species.Description = str_to_lower(Species.Description), #removes capitializations
Species.Description = str_replace_all(Species.Description, "-", " "), #removes dashes
Animal.Class = str_to_lower(Animal.Class),
Animal.Class = case_when(
str_detect(Animal.Class, "marine reptiles") ~ "terrestrial reptile or amphibian", #moving turtles to correct category
str_detect(Animal.Class, "marine mammals") ~ "marine mammals",
str_detect(Animal.Class, "domestic") ~ "domestic",
str_detect(Animal.Class, "raptors") ~ "raptors",
str_detect(Animal.Class, "fish") ~ "fish",
TRUE ~ Animal.Class),
Species.Description = case_when(
str_detect(Species.Description, "squirrel") ~ "squirrel", #more examples of this but this was most obvious
TRUE ~ Species.Description
)
)
write.csv(data_clean, "UrbanRangerDataClean.csv")
Overall, this is much better now and solved some issues. I consolidated the Animal.Class category to make further analysis easier, but recognize this may have over simplified some details. There are some cases when some species are labelled different ways. We’ll leave this for now.
#Grouping by species for top 20 most abundant species
species <- data_clean %>%
count(Species.Description, sort = TRUE) %>%
top_n(20)
## Selecting by n
#Species count
plot <- ggplot(data = species, aes(x= reorder(Species.Description, n), y = n)) + #reorder makes bars descending order
geom_bar(stat = "identity") + #allows for the count to be plotted
coord_flip() + #rotates graph
labs(title= "Top 20 Species Count", y= "Count",x ="Species") +
theme_minimal() #introduces a theme to the figure instead of the standard output
plot
Funny to see that chickens are the fourth most reported animal in this dataset!
With such a high abundance of calls for raccoons, I’m going to remove them from the next graphic. This way we can understand how the rest of the species distributions look. Looking at how the Animal.Class reported across the boroughs.
plot2 <- data_clean %>%
filter(Species.Description != "raccoon") %>%
group_by(Borough) %>%
count(Animal.Class, sort = TRUE)%>%
ggplot(aes(x= reorder(Animal.Class, n), y = n)) + #reorder makes bars descending order, n represents count
geom_bar(aes(fill=Borough), stat = "identity") + #allows for the count to be plotted
scale_fill_brewer(palette = "Greens") +
theme_classic()+ #introduces a theme to the figure instead of the standard output
facet_wrap(~Borough, nrow=1)+
labs(title = "Reports by Animal Classification, without raccoons", y= "Number of individuals reported", x=NULL)+
coord_flip()
plot2
Now that we know that there are differences across the boroughs, let’s take a look at the most popular places animals are reported. Not surprising, this dataset falls within the NYC Parks most often.
plot3 <- data_clean %>%
count(Property, Borough, sort = TRUE) %>%
top_n(10, n) %>%
ggplot(aes(x=reorder(Property, -n), y=n)) +
geom_bar(aes(fill = Borough), stat = "identity")+
scale_fill_brewer(palette = "Greens")+
xlab("")+
ylab("Number of reported animals")+
theme_minimal()+
theme(axis.text.x = element_text(angle = 45, hjust = 1)) #rotates labels 45 degrees and adjust down to not overlap axis
plot3
