Temperature variation plot in ggplot

Question

I have hourly temperature data from an old experiment. I would like to summarize the dataset into a graph in ggplot showing the mean temperature of each experimental treatment, and the variation in temperatures within each treatment every hour. I would like to have a non-excel graph that looks something like this:

The data are linked here. https://www.dropbox.com/sh/27cft3118dha6xt/46_xxZZano

I probably have to use reshape to format the data correctly. JD refers to the Julian Day of the year and Time is the hour within that day. Note that labels A-H are treatment 1, I –P are treatment 2 and Q-X are treatment 3. Any advice on how to best go about this would be greatly appreciated.

Many thanks.

Answer 1

Maybe like this:

df<-read.csv(file="2011_Temps_obs.csv")

require(reshape2)
require(ggplot2)
require(dplyr) # for aggregation

df$Hour<-as.character(                          # load "Hour" in correct format
  strptime(
    paste(df$Year,df$JD,substr(
      formatC(df$Time, width = 4, format = "d", flag = "0"),
      1,2),sep="-"),
    format="%Y-%j-%H"
  ))

m<-melt(df,id.vars="Hour")                      # melt by hour
m<-m[!(m$variable %in% c("Year","JD","Time")),] # filter out un-needed columns

lookup<-data.frame(variable=unique(m$variable),test=c(rep(1,5),rep(2,8),rep(3,25)))

ggplot(merge(m,lookup,by="variable")) +         # merge m to get the test rollup
  geom_smooth(aes(x=Hour,y=value,group=as.factor(test),fill=as.factor(test),color=as.factor(test)))

This is the smoothed graph with CIs

OR this pre-calculating your own summary stats using dplyr

summdata<-
  merge(m,lookup,by="variable") %.%
  group_by(Hour,test) %.%
  summarise(mean=mean(value),min=min(value),max=max(value))

ggplot(summdata,aes(group=as.factor(test), color=as.factor(test), fill=as.factor(test))) +         
  geom_line(aes(x=Hour,y=mean),size=1,alpha=0.6) + 
  geom_ribbon(aes(x=Hour,ymin=min,ymax=max),alpha=0.1)