Reading NASA global temperature index using R

read_gistemp.R

### Download file
url <- "http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt"
tmp <- tempfile()
download.file(url, destfile = tmp, method = "curl")

### Read the file and select colnames
dat <- readLines(tmp)
col_names <- unique(grep("^Year", dat, value = TRUE)) ## select colnames
col_names <- strsplit(col_names, "\\s+")[[1]]

### Remove header and turn '*' to NA
dat <- grep("^\\d+", dat, value = TRUE) ### select lines that start with a number (year)
dat <- gsub("\\*+", "NA", dat) ### *** == NA

### Collapse it to a unique character to and read it using read.table
dat <- paste(dat, collapse = "\n")
d <- read.table(text = dat, header = FALSE, col.names = col_names)
str(d)
## 'data.frame':	135 obs. of  20 variables:
##  $ Year  : int  1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 ...
##  $ Jan   : int  -32 -12 5 -36 -18 -55 -35 -57 -41 -17 ...
##  $ Feb   : int  -27 -15 6 -36 -12 -27 -41 -41 -39 17 ...
##  $ Mar   : int  -22 -1 0 -11 -29 -17 -32 -24 -37 8 ...
##  $ Apr   : int  -29 -1 -22 -18 -34 -35 -21 -32 -23 8 ...
##  $ May   : int  -16 -1 -18 -18 -31 -33 -19 -26 -24 0 ...
##  $ Jun   : int  -23 -25 -30 -6 -34 -38 -28 -23 -19 -9 ...
##  $ Jul   : int  -18 -10 -25 -1 -30 -26 -11 -17 -9 -11 ...
##  $ Aug   : int  -10 -6 -9 -11 -23 -22 -19 -26 -12 -17 ...
##  $ Sep   : int  -19 -16 -9 -17 -27 -16 -10 -22 -9 -17 ...
##  $ Oct   : int  -18 -21 -23 -17 -23 -12 -21 -31 -2 -20 ...
##  $ Nov   : int  -15 -26 -23 -26 -27 -12 -27 -27 0 -31 ...
##  $ Dec   : int  -21 -16 -35 -18 -24 2 -15 -38 -9 -28 ...
##  $ J.D   : int  -21 -12 -15 -18 -26 -24 -23 -30 -19 -10 ...
##  $ D.N   : int  NA -13 -14 -19 -26 -26 -22 -28 -21 -8 ...
##  $ DJF   : int  NA -16 -2 -36 -16 -35 -25 -38 -39 -3 ...
##  $ MAM   : int  -22 -1 -13 -16 -31 -28 -24 -27 -28 6 ...
##  $ JJA   : int  -17 -13 -21 -6 -29 -29 -19 -22 -13 -12 ...
##  $ SON   : int  -18 -21 -18 -20 -26 -13 -19 -26 -4 -22 ...
##  $ Year.1: int  1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 ...



### Bonus 
library(dplyr)
library(tidyr)
library(ggplot2)

d <- d %>%
  select(Year, Jan:Dec) %>%
  gather(month, temp, Jan:Dec) %>%
  mutate(date = paste(Year, month, "01", sep = "-"),
         date = as.Date(date, "%Y-%b-%d")) %>%
           select(date, temp)

ggplot(d, aes(date, temp)) +
  geom_point(shape = 19, size = 0.5) +
  geom_line(size = 0.2) +
  geom_smooth(colour = "red")