-1

I am trying to import some .csv files into R for a company. I am trying to basically compare SPC over a selected date domain. When I enter the data into R I have set all other columns to NULL and then I try to plot .csv file. When I do this, my dates are out of order, I have used sapply(mydata2, class) and find that Date is a factor and SPC is an integer. I am sure this is part of the problem and has been consistently the issue. I have slightly remedied the situation by changing the Excel file (which is a CSV file) date column to a Julian date but for sake of presentation I would much rather have it in short date format. Also it would be great to know how to do this in R as opposed to having to switch to Excel. This has also been my issue for kmeans clustering as well.

Any ideas?

I should also mention that I am basically trying to create a simple function that removes all the nonsense from the Excel file and basically computes the necessary components in various ways. I have roughly 60 more Excel files, split by month to perform this analysis on.

mydata2 = read.csv("Copy of Monthly Raw SPC Aug 2015.csv")
mydata2
mydata2$Trailer <- NULL
mydata2$ProducerID <- NULL
mydata2$SampleID <- NULL
mydata2$Producer.Number <- NULL
mydata2$BTUNo <- NULL
mydata2$Route <- NULL
mydata2
plot(mydata2)
sapply(mydata2, class)

It is just a simple code for plots, I have tried other things like ordering or boxplots. A sample of the actual data I want to plot is, 

...
96  42233  27000
97  42233  29000
98  42233   2000
99  42233  38000
100 42234  11000
101 42234 157000
...

Instead of the general number, first column would be in a short date formate like 96 would be 8/16/2015. So what happens is when I then go to plot, my box and whisker plot has multiple entries on the same day but the days are all over the graph, I need the same result but the days ordered.

Vincent
  • 1
  • 1
  • 2
    Hi, welcome to SO. Please provide a [reproducible example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) including your code and some of your data. – Heroka Oct 07 '15 at 14:44
  • your date is a factor, which will not be ordered sensibly. You want to use the function `as.Date` on it, but without sample data, that's all we can tell. – jeremycg Oct 07 '15 at 14:45
  • I should also add that yeah, you are right, when I run sapply when I have dates they show up as factors, but when I switch them to a julian date I think it shows up as an integer or something. Then I can plot it. I want to be able to order dates in plots as well as in plotting kmeans. – Vincent Oct 07 '15 at 19:21

1 Answers1

0

Well I figured it out, after some tweeking here it is, 

 library(chron)
 library(ggplot2)
 library(maptools)
 mydataE = read.csv("Copy of Monthly Raw SPC Aug 2015.csv")
 mydataE
 mydataRAW <- mydataE[-(161:163),]
 attach(mydataRAW)
 Date4 <- as.vector(Date)
 Date4
 Date6 <- dates(Date4, format = c(dates ="m/d/Y"), out.format = c(dates= "d/m/Y")) 
 Date6
 u <- ggplot(dfbest, aes(factor(Date6),SPC))
 u + geom_boxplot()

So there it is, it was able to control how I entered the dates and thus order the x-axis, before, my dates would read as, 8/10/2015...8/20/2015...8/31/2015...8/4/2015...8/5/2015... then it would display them on my x-axis as such, so now I have total control on how the x-axis is arranged and the graphs look better. Also I removed some of the bad syntax, or at least it looked that way, it seemed that when I would NULL the multiple columns that I did not need, and just plot Date with SPC it would give the boxplot that I wanted, but now with attach() it works provided I use chron and ggplot2.

Vincent
  • 1
  • 1