How to plot data in R from different days overlayed in only one day (based on hours:minutes)?

Question

I am having this problem with R:

I have a dataset called "teste" that has a column for 'Date' (it's in POSIXct, format = "%Y-%m-%d %H:%M:%S"), which has readings every 10 minutes during a 5 months period.

I need to make a comparison of the variables at the same time but in different days. For example plot every Saturday in the Dataset, overlayed. I already have the code for subsetting the data.frame and have only the Saturdays.

Here is a sample of the data:

  DATE                ID   VAR1   VAR2  VAR3
1 2016-09-19 00:07:47 79    19     0    OPN
2 2016-09-19 00:17:47 79    18     1    OPN
3 2016-09-19 00:27:47 79    16     3    OPN
4 2016-09-19 00:37:47 79    15     4    OPN
5 2016-09-19 00:47:47 79    16     3    OPN
6 2016-09-19 00:57:47 79    16     3    OPN

Here is the dput from the data:

structure(list(FECHA = structure(c(1474236467, 1474237067, 1474237667, 1474841253, 1474841853, 1474842453), class = c("POSIXct", "POSIXt" ), tzone = ""), ID = c(79L, 79L, 79L, 79L, 79L, 79L), SLOTS = c(19L, 18L, 16L, 14L, 15L, 15L), BIKES = c(0L, 1L, 3L, 8L, 7L, 7L), STATUS = structure(c(2L, 2L, 2L, 2L, 2L, 2L), .Label = c("CLS", "OPN"), class = "factor")), .Names = c("FECHA", "ID", "SLOTS", "BIKES", "STATUS"), row.names = c(1L, 2L, 3L, 1004L, 1005L, 1006L ), class = "data.frame")

I tried doing: (used package lubridate)

plot(as.POSIXct(paste(hour(teste$FECHA),":",minute(teste$FECHA),sep = ""), format = "%H:%M"),teste$BIKES)

It works but using 'paste' is definetely not the best way for doing this. There probably is a easier and more elegant way, right? If yes, how?

And I have a problem if I plot it with type = "lines", because it doesnt know that the last reading of a day shouldnt be 'connected' to the first of the next one, giving this result: (see the lines crossing all the graph from 24 hours to 0 hours)

I though about ploting one day at a time, using the plot and then using lines functions, but the problem is that I dont know how many readings are there in each day. Each day should have 6*24=144 readings, but some have 143, 142 (because of problems while getting data).

I appreciate any help.

you can try `library(ggplot2); ggplot(data, aes(x=DATE, y=VAR1)) + geom_point() + geom_line()` — Roman, Apr 19 '17 at 08:54
@Jimbou I tried and it plots as different days, I'd like to have all plots in the same 0 hours till 24 hours scale, even though the data is from different dates. This is what I got: http://i.imgur.com/R8tFadg.png — Bruno Caram Muller, Apr 19 '17 at 09:00
Then please edit your reproducible example data accordingly. You can use `dput(subset_of_your_data)` and paste the output in your question. — Roman, Apr 19 '17 at 09:17
ops, sorry. Here is what I get for 2 fragments (first 3 readings) of 2 different days: structure(list(FECHA = structure(c(1474236467, 1474237067, 1474237667, 1474841253, 1474841853, 1474842453), class = c("POSIXct", "POSIXt" ), tzone = ""), ID = c(79L, 79L, 79L, 79L, 79L, 79L), SLOTS = c(19L, 18L, 16L, 14L, 15L, 15L), BIKES = c(0L, 1L, 3L, 8L, 7L, 7L), STATUS = structure(c(2L, 2L, 2L, 2L, 2L, 2L), .Label = c("CLS", "OPN"), class = "factor")), .Names = c("FECHA", "ID", "SLOTS", "BIKES", "STATUS"), row.names = c(1L, 2L, 3L, 1004L, 1005L, 1006L ), class = "data.frame") — Bruno Caram Muller, Apr 19 '17 at 10:16

score 0 · Answer 1 · answered Apr 19 '17 at 11:04

You can try

# add factor to differentiate between the different days
d$day <- factor(strftime(d$FECHA, format="%Y-%m-%d"))
# add hours, set one day
d$time <- as.POSIXct(strftime(d$FECHA, format="%H:%M:%S"), format="%H:%M:%S")
# color per day
ggplot(d, aes(x=time, y=BIKES, col=day)) + 
       geom_point() + 
       geom_line() + 
       scale_x_datetime(limits = as.POSIXct(c("2017-04-19 00:00:00", "2017-04-19 24:00:00")), date_labels = "%H:%M")
# or facet_wrap/grid
ggplot(d, aes(x=time, y=BIKES)) + geom_point() + geom_line() + 
  facet_grid(~day)+
  scale_x_datetime(limits = as.POSIXct(c("2017-04-19 00:00:00", "2017-04-19 24:00:00")), date_labels = "%H:%M")

# or use the wrap function instead of the grid function
  facet_wrap(~day, scales = "free_y")

TichPi · Answer 2 · 2022-02-02T13:11:28.450

Please try this:

#Generate dataset
teste=structure(list(FECHA = structure(c(1474236467, 1474237067, 1474237667, 1474841253, 1474841853, 1474842453), class = c("POSIXct", "POSIXt" ), tzone = ""), ID = c(79L, 79L, 79L, 79L, 79L, 79L), SLOTS = c(19L, 18L, 16L, 14L, 15L, 15L), BIKES = c(0L, 1L, 3L, 8L, 7L, 7L), STATUS = structure(c(2L, 2L, 2L, 2L, 2L, 2L), .Label = c("CLS", "OPN"), class = "factor")), .Names = c("FECHA", "ID", "SLOTS", "BIKES", "STATUS"), row.names = c(1L, 2L, 3L, 1004L, 1005L, 1006L ), class = "data.frame")
    
#marked by date
teste$day <- format(teste$FECHA,'%Y-%m-%d')
   
#marked by hour & minute
teste$hm <- format(teste$FECHA, format = "%H:%M")
    
#plot it with hour & minute as x, and day as group
library(ggplot2)
ggplot(teste, aes(x=hm, y=BIKES, group=day)) + geom_point() + geom_line()

And if you want this plot more elegent, your can decorate it with ggplot2 grammer. For example:

ggplot(teste, aes(x=hm, y=BIKES, group=day)) + 
geom_point(aes(colour = factor(day))) + 
geom_line(aes(colour = factor(day))) +
theme_bw()+labs(colour="Date")

How to plot data in R from different days overlayed in only one day (based on hours:minutes)?

2 Answers2

Linked