land_temps.R

library(readr)
library(dplyr)
library(ggplot2)
library(janitor)
library(broom)
library(lubridate)

# available at https://www.kaggle.com/berkeleyearth/climate-change-earth-surface-temperature-data
temp_loc <- "/Users/conormcdonald/Downloads/GlobalLandTemperatures/GlobalLandTemperaturesByCountry.csv"

temp <- read_csv(temp_loc)

temp <- temp %>% 
  clean_names() %>% 
  filter(!is.na(averagetemperature)) %>% 
  mutate(year = year(dt), 
         month = month(dt))

country_fit <- temp %>% 
  group_by(country) %>% 
  do(glance(lm(averagetemperature ~ dt, data = .)))

country_fit %>% 
  ungroup() %>% 
  filter(r.squared > 0.10) %>% 
  mutate(country = reorder(country, r.squared)) %>% 
  ggplot(aes(country, r.squared, colour = r.squared)) + 
  geom_point(alpha = 0.8) + 
  coord_flip() +
  guides(colour = FALSE) +
  labs(title = "R-Sqaured of average land temperature explained by time",
       subtitle = "Showing countries with R-Sq > 0.10", 
       y = "R-Sqaured", 
       x = "Country") 

temp %>% 
  filter(country == "Indonesia") %>% 
  group_by(year) %>% 
  summarise(ave_temp = mean(averagetemperature)) %>% 
  ggplot(aes(year, ave_temp)) +
  geom_point(alpha = 0.6) +
  geom_smooth(method = "lm") +
  labs(title = "Average land temperature by year for Indonesia")