Using geom_rect with a time series in POSIXlt

Question

I am using ggplot2 to make a simple two variable plot. My data set "Downstream" contains data from a downstream salinity probe and data from a hand-held conductivity meeter. I am using this plot to show the large amount of data missed by the hand-held conductivity meter. I hope to highlight a few of the peaks in the data missed by the handheld conductivity meter using the geom_rect layer. So far my bulky code looks like this:

library(ggplot2)
> Downstream <- read.csv("~/Desktop/Downstream.csv")
> Date= as.character(Downstream$Date)
> Date=strptime(Date,format=("%m/%d/%y %H:%M"))
> Downstream$Date=Date
> Dplot <- ggplot(data= Downstream,aes(x=Date))
> Dplot <- Dplot + geom_line(aes (y=Conductivity), color="blue")
> Dplot <- Dplot + geom_point(aes(y=Thermo.Conductivity),color="red")
> Dplot<- Dplot + ggtitle("Logger and Hand Sample \nConductivity vs. Time ") +
+ theme(plot.title = element_text(lineheight=.8, face="bold"))
> Dplot <- Dplot + ylim(0,4000)

This returns this plot:

I am reasonably happy with this. All I have left to do is add the geom_rect layer, but this has proven a challenge. The head of my dataset looks like this:

                 Date Water.Level Conductivity Thermo.Conductivity
1 2013-12-17 22:00:00       0.216       487.79                  NA
2 2013-12-17 22:15:00       0.210       487.38                  NA
3 2013-12-17 22:30:00       0.220       485.77                  NA
4 2013-12-17 22:45:00       0.225       485.37                  NA
5 2013-12-17 23:00:00       0.236       484.96                  NA
6 2013-12-17 23:15:00       0.241       486.19                  NA

The structure of my data set looks like this:

'data.frame':   23472 obs. of  4 variables:
 $ Date               : POSIXlt, format: "2013-12-17 22:00:00" "2013-12-17      22:15:00" ...
 $ Water.Level        : num  0.216 0.21 0.22 0.225 0.236 0.241 0.238 0.231     0.217 0.235 ...
 $ Conductivity       : num  488 487 486 485 485 ...
 $ Thermo.Conductivity: num  NA NA NA NA NA NA NA NA NA NA ...

The closest solution to the problem that I'm having that I have seen is posted here: link

I would like to end up with something like this but I have been unable to utilize the solutions presented.I think my problem is that "Date" is a POSIXlt and not a POSIXct. Resent (and embarrassing) failures include:

Dplot<- Dplot+ annotate(geom_rect(),x=Date,y=Conductivity,xmin= 2014-01-03     04:15,xmax=2014-01-05 12:30,ymin=-Inf,ymax=Inf,)
Error: unexpected numeric constant in "Dplot<- Dplot+     annotate(geom_rect(),x=Date,y=Conductivity,xmin= 2014-01-03 04"

Dplot<- Dplot+ annotate("rect",fill="gray",alpha(=.5),xmin= 2014-01-03     04:15,xmax=2014-01-05 12:30,ymin=-Inf,ymax=Inf,)
Error: unexpected '=' in "Dplot<- Dplot+ annotate("rect",fill="gray",alpha(="

d.water<- data.frame(x1=c(2014-01-03 04:15:00,2014-02-18 11:45:00,2014-03-17     12:15:00,2014-05-14 18:15:00),x2=c(2014-01-05 12:30:00,2014-02-20 3:30:00,2014-03-    21 14:30:00,2014-05-16 05:15:00),y1=c(-Inf,-Inf,-Inf,-Inf),y2=c(Inf,Inf,Inf,Inf))
Error: unexpected numeric constant in "d.water<- data.frame(x1=c(2014-01-03 04"

Answer 1

Here is what I would do.

First generate a new data.frame that specifies your bounds for your rectangle:

rect_df <- data.frame(x1=c("2014-01-03 04:15:00","2014-02-18 11:45:00","2014-03-17 12:15:00","2014-05-14 18:15:00"),x2=c("2014-01-05 12:30:00","2014-02-20 3:30:00","2014-03-21 14:30:00","2014-05-16 05:15:00"),y1=c(-Inf,-Inf,-Inf,-Inf),y2=c(Inf,Inf,Inf,Inf), DataMatch=c("First","Second","Third","Fourth"))

rect_df$x1=strptime(rect_df$x1,format=("%Y-%m-%d %H:%M:%S"))
rect_df$x2=strptime(rect_df$x2,format=("%Y-%m-%d %H:%M:%S"))

rect_df

Then add it to your plot like this:

Dplot +
  geom_rect(data=rect_df, aes(xmin=x1, yxmax=x2, ymin=y1, ymax=y2, fill=DataMatch, alpha=0.5))

What you are doing here is asking ggplot2 to reference a different data.frame that the original one for this particular geom. The trick here is that the dates need to be in the same format. Without a reproducible example, it isn't possible to test whether this works. If you are serious about using R, I'd recommend carefully reading this thread to learn how ask good questions: How to make a great R reproducible example?