Rlang

• options(error=recover) which drops you into a nice lispy stack

• devtools - package for simplifying working with packages

• ggplot flipbook

• Set the default timezone to UTC

  Sys.setenv(TZ='GMT')

• Plots vs time

  require(ggplot2) require(scales) ggplot(subset(x, time>'2017-05-15'), aes(x=time, y=updates)) + geom_line()+scale_x_datetime(breaks = date_breaks('1 day'), minor_breaks = date_breaks('6 hour'))

• Create date from numbers - use ISOdate or datetime from numbers - use ISOdatetime

• Parse time line Jul 18, 2016 11:43:10 PM into datetime object

  x$time<-as.POSIXct(x$V1, format='%b %d, %Y %I:%M:%S %p')

  (see here for a function that automatically determines the format and here for information on POSIXct)

• Parse epoch (seconds!) into time

  x$time <- as.POSIXct(x$V1, origin="1970-01-01") # may add tz='GMT'

• ggplot tutorial

• Group by date and sum

  require(dplyr) x$time <- as.POSIXct(x$time) x$date <- as.Date(x$time) x %>% group_by(date) %>% summarise(size = sum(downloadSize)/1e6) # use =n() for counting

• Set ylimit, xlimit (effectively zoom)

  • coord_cartesian(ylim=c(-12,12))

Interesting resources

• blog