Timestamp looks correct but appears wrong on ggplot2

Question

I have a time series of CO2 data with a UNIX timestamp that looks like this: 1658759863

e.g.

df <- data.frame(timestamp=seq(1653998631,1663998631,10),co2_ppm=runif(1000001))

I then convert to POSIXct and filter data only after 11:52:51

    df <- df%>% 
  mutate(timestamp=as.POSIXct(timestamp,origin="1970-01-01",tz = "GMT"))%>%
  filter(timestamp>"2022-07-28 11:52:51" )

attr(df$timestamp,"tzone")

[1] "GMT"

When I filter it to only plot value after "11:52:00", it returns after "10:52:00"

why?

Seems like a timezone issue to me. Your timestamps are defined as GMT, what is your local timezone dplyr::filter() seems to make use of per default? — falk-env

Rui Barradas Rui Barradas · Accepted Answer · 2022-07-28T15:10:09

This is a daylight savings issue, see the example below.

s <- Sys.time()
as.POSIXct(as.integer(s), origin = "1970-01-01", tz = "GMT")
#> [1] "2022-07-28 15:05:57 GMT"
as.POSIXct(as.integer(s), origin = "1970-01-01", tz = "Europe/London")
#> [1] "2022-07-28 16:05:57 WEST"
s
#> [1] "2022-07-28 16:05:57 BST"

^{Created on 2022-07-28 by the reprex package (v2.0.1)}

R's time is the number of seconds since an origin and time zone and though London and Greenwich are in the same time zone, London's official time is off by 1 hour in this time of year.

Timestamp looks correct but appears wrong on ggplot2

1 Answers