Temperature of every hour in London since 1973

hourly-weather-london.R

library(tidyverse)
library(R.utils)
library(glue)
library(padr)
library(ggExtra)
library(lubridate)
library(scales)
library(hrbrthemes)

#london
USAF <- '037720'
WBAN <- '99999'

url <- 'ftp://ftp.ncdc.noaa.gov/pub/data/noaa/isd-lite/{yr}/{usaf}-{wban}-{yr}.gz' 

urls <- glue(url, yr = 1973:2017, usaf = USAF, wban = WBAN)

fetch <- safely(function(url, .pb = NULL) {
  if ((!is.null(.pb)) && inherits(.pb, "Progress") && (.pb$i < .pb$n)) .pb$tick()$print()
  
  url %>%
    downloadFile() %>%
    gunzip(tempfile()) %>%
    read_fwf(
      col_positions = fwf_cols(
        year = c(1, 4),
        month = c(6, 7),
        day = c(9, 11),
        hour = c(12, 13),
        temp = c(14, 19)
      ), 
      col_types = "iiiid"
    )
})
pb <- progress_estimated(length(urls))
weather_data <- map(urls, fetch, .pb = pb)

data <- weather_data %>% 
  map_df(~.x$result) %>% 
  filter(temp != -9999) %>%
  mutate(temp = temp/10) %>% # downloaded data is scaled by a factor of 10
  mutate(timestamp = make_datetime(year, month, day, hour)) %>%
  # add and fill some missing hours here and there
  pad() %>%
  mutate(
    year = year(timestamp),
    month = month(timestamp),
    day = mday(timestamp),
    hour = hour(timestamp)
  ) %>%
  fill(temp) 

data %>%
  mutate(idx = (yday(timestamp)-1)*24 + hour) %>% # creates hourly index

  # the following block removes leap dates and readjusts hourly index:
  mutate(
    idx = if_else(
      leap_year(year) & idx > 1416,
      idx - 24,
      idx
    )
  ) %>%
  filter(!(month == 2 & day == 29)) %>% 

  ggplot(aes(x = idx, y = year, fill = temp)) + 
  geom_tile(height = .9) +
  scale_y_reverse(
    expand_scale(add = c(1, 0)),
    breaks = 1973:2017
  ) +
  scale_x_continuous(
    expand = c(0,0),
    breaks = c(0, 1094, 2189, 3284, 4379, 5474, 6569, 7664, 8759),
    labels = c(
      "1 Jan, 12:00", 
      "15 Feb, 14:00",
      "2 Apr, 5:00", 
      "17 May, 20:00",
      "2 Jul, 11:00",
      "17 Aug, 2:00",
      "1 Oct, 17:00",
      "16 Nov, 8:00",
      "31 Dec, 23:00"
    )
  ) +
   scale_fill_viridis_c(
    breaks = c(-11, 1, 13, 25, 37),
    labels = c("-11℃", "1℃", "13℃", "25℃", "37℃"), # data range is -11 - 37.3
    option = "C"
  ) +
  theme_ipsum() +
  labs(
    x = "Hour of the year",
    y = "Year",
    fill = "Temp",
    title = "Hourly Temperature in London", 
    subtitle = "1973-2017",
    caption = "Source: NCEI/NOAA, www.ncdc.noaa.gov"
  )


ggsave("~/Downloads/weather/weather-london-hi-res.png", scale = 9, limitsize = F)
ggsave("~/Downloads/weather/weather-london.png", scale = 3)