Wine PhD

Northern California (and the rest of CA, too) is in the midst of a significant rain event. I thought things were wet in Davis. Oakville, in the heart of Napa Vally, has received 3-4 times as much rain, so far.

The somewhat circuitous split/do.call/rbind was necessary to get around an issue with the groups argument to lattice, which would not do what I wanted. Good explanation of my problem, no?

#R code
library(cimis)
library(lattice)
#northern ca weather

w <- mcimishourly(c("006", "077"), "name")

xyplot(do.call("c", lapply(split(w, w$station),
	function(x)cumsum(x$precip))) ~ w$datetime,
	groups = w$station, type = "l",
	auto.key=T, xlab = "Time",
	ylab = "Cumulative Precipitation, in")

I find that I use my cimis package almost every day to check the weather trends in Davis. It's pretty easy to do:

library(cimis)
xyplot(air_temp ~ datetime, data = w <- cimishourly('006'), type = 'l')

Things get more complicated when I want to compare multiple stations or multiple years (with cimisannual). This was a the major feedback I got from the one person I know who uses this package (thanks Kim!).

In the updated version, I have added two new functions: mcimishourly and mcimisannual, which take multiple stations and years as function parameters.

Below is a graph comparing 9 years worth of air temperature data for 4 weather stations. Granted, it's not very useful, but it does give a good overview of the trends and differences between the sites.

Below is the code to create the graph. Most of the code is for graphing the data. The stations and years could be written inline as one line of code.

#code for new cimis 0.1-3 updates
library(lattice)
library(cimis)

places <-c('006', '077', '075', '052')
years = 2000:2008

w <- mcimisannual(places, years, label = "name")

xyplot(avg_air ~ julian_day | as.factor(year),
		groups = station, type  = 'l', data = w,
		layout = c(1,9), strip.left=T, strip=F,
		auto.key=list(columns=2), xlab = "Julian Day",
		ylab = "Average air temp, F")

I also fixed the leading white spaces in the stnlist, which isn't so important except to make the legend labels align properly.

As always, any feedback would be greatly appreciated.
I'm uploading the new version tonight, hopefully it will be on CRAN in the next few days. Until then,
you can get a copy here: cimis_0.1-3.tar.gz. Now available on CRAN at http://cran.r-project.org/web/packages/cimis/index.html or use

install.packages("cimis")