Preprocessing and simple display of launches

launchlog.R

####Pre-processing of launch log before sending it to Google Refine####
#The vector of widths can be grabbed a number of ways: by hand, with regular expressions, or from a file explaining the structure
launch.width<-c(12,27,15,31,26,9,24,16,9,12,5,19)
#Your file path will vary, of course
launch.path<-c("/Users/protonk/Dropbox/Past Econ Work/Launch vehicle market/Complete text launch log.txt")
launch.names<- c("Launch","Date","COSPAR","PL.Name","Orig.PL.Name","SATCAT","LV.Type","LV.SN","Site","Subsite","Success","Ref")
#Read in FWF and export a .csv
launch.test<-read.fwf(launch.path,launch.width,header=FALSE,skip=2,as.is=TRUE,col.names=launch.names)
write.csv(launch.test,file="~/R/launch.csv")
##########
##Google Refine Magick
##########
library(ggplot2)
library(plyr);
########Processing of the exported file
launch.trimmed<-read.csv("~/Downloads/launch.csv",as.is=TRUE)
#This is just to make looking at summary and structure easier
launch.min<-subset(launch.trimmed,select=c(Launch,Date,Year,Orig.PL.Name,Site,Success,Lat,Long))
#Note to self, don't make year a factor (year was a variable I created in Google Refine)
launch.min$Site<-as.factor(launch.min$Site)
launch.min$Date<-as.Date(launch.min$Date,"%Y %b %d")
#The dataset is pretty much complete, the NA's are piggyback launches
launch.min<-launch.min[which(complete.cases(launch.min)),]
names(launch.min)<-tolower(names(launch.min))
#By year/site summary stats
launch.sum<-ddply(launch.min,c("site","year"),"nrow")
names(launch.sum)<-c("site","year","launches")
#The plot
launch.plot<-qplot(year,launches,data=launch.sum,group=site,colour=site,geom="line")+opts(legend.position = "none")+scale_y_continuous(name="Launches per Year")+scale_x_continuous(name='')