In sophiasternberg/sspalmer:

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(sspalmer)
library(ggplot2)

head(pal_penguin)

# cleaning extraneous values
pal_penguin <- pal_penguin[!(pal_penguin$`Mean Mass Flux (mg/m²/day)` < 0),] 

g <- ggplot(data = pal_penguin, aes(x = `Start Date`, y = `Mean Mass Flux (mg/m²/day)`)) +
  geom_point() +
  labs(x = "Date",
    y = "Mean Mass Flux (mg/m²/day)",
    title = "Mean Mass Flux Overtime")  +
  theme(plot.title = element_text(hjust = 0.5))
g

g <- ggplot(data.frame(pal_penguin$`Mean Mass Flux (mg/m²/day)`), aes(pal_penguin$`Mean Mass Flux (mg/m²/day)`)) +
  geom_histogram(binwidth = 60, aes(y = ..density..))  +
  labs(x = "Mean Mass Flux (mg/m²/day)",
    y = "Density")  +
  theme(plot.title = element_text(hjust = 0.5))
g

g + geom_density(bw = 20, kernel = "gaussian", col = "red")

g + geom_density(bw = 20, kernel = "rectangular", col = "dodgerblue")

g + geom_density(bw = 20, kernel = "epanechnikov", col = "blue")

g + geom_density(bw = "ucv", kernel = "gaussian", col = "pink") +
  geom_density(bw = "nrd", kernel = "gaussian", col = "purple")

Out of all the density curves, it looks like a bandwidth of 20 with the epanechnikov kernel fits the Mean Mass Flux data the best.

sophiasternberg/sspalmer documentation built on Dec. 23, 2021, 4:22 a.m.