Our assignment this week was to debug a block of code.

The original code was:

tukey_multiple <- function(x) {
   outliers <- array(TRUE,dim=dim(x))
   for (j in 1:ncol(x))
    {
    outliers[,j] <- outliers[,j] && tukey.outlier(x[,j])
    }
outlier.vec <- vector(length=nrow(x))
    for (i in 1:nrow(x))
    { outlier.vec[i] <- all(outliers[i,]) } return(outlier.vec) }

To start debugging, I pasted the code into RStudio to see what kind of error it would throw. I got:

Error: unexpected symbol in:
"    for (i in 1:nrow(x))
    { outlier.vec[i] <- all(outliers[i,]) } return"

The “unexpected symbol” error made me think that some brackets or parentheses weren’t closed or were misplaced. That wasn’t it, but it turned out to be something similar – looking at the last line, I noticed that the “return” looked out of place. It should have been on a separate line, by itself. That would certainly make it an unexpected symbol. I moved return to the next line, making the code look like:

tukey_multiple <- function(x) {
  outliers <- array(TRUE,dim=dim(x))
  for (j in 1:ncol(x))
  {
    outliers[,j] <- outliers[,j] && tukey.outlier(x[,j])
  }
  outlier.vec <- vector(length=nrow(x))
  for (i in 1:nrow(x))
  { outlier.vec[i] <- all(outliers[i,]) }
  return(outlier.vec) }

At that point, the code ran successfully and the function was created.

Using the function was a different story – I passed it a built-in list as an argument, and got a new error:

Error in tukey.outlier(x[, j]) : could not find function "tukey.outlier"

I’m guessing that tukey.outlier() is a function that returns TRUE or FALSE depending on whether the value is an outlier. The code runs fine without that function (albeit it makes tukey.multiple() always return a bunch of TRUEs), so I assume this code will work provided tukey.outlier() is defined.

This week’s assignment was to create a DESCRIPTION file for our final project package.

The DESCRIPTION file I created is on GitHub here:

https://github.com/jered0/lis4930-rpackage/blob/master/queryNIBRS/DESCRIPTION

Getting the DESCRIPTION format right was interesting, and I ended up looking at a few packages on CRAN to see what package names typically look like and how they describe themselves. For example, I thought at first that the Title field was just an expanded name for the package, but I see that it does, as you said in the module, act more as a descriptive subtitle.

Because my package will involve retrieving data from the FBI’s Crime Data API, I added an Imports section for the dependency the package will need, RCurl. I’ll look into whether that’s strictly required – I might be able to use R’s built-in web queries instead of relying on RCurl’s fancy capabilities.

This week the assignment was to plot graphs for a data set using a built-in R function, a function from the lattice package, and a function from the ggplot2 package.

The source code I created for the task can be found on GitHub here:

https://github.com/jered0/lis4930-rpackage/blob/master/module9/lattice-ggplotting.r

I used the SeaSlugs data set. A description of the set is on this page, and the data set itself can be downloaded via this link.

Given that there were two categories of data (time and metamorphose percentage), and there were multiple percentages for each time entry, I decided to represent the data with a boxplot.

Built-in function

I used the built-in boxplot() function to create the first graph.

I’ve used this function before, but I did have to fiddle with it to work out how to get the y-axis labels horizontal (I thought it looked better that way), and to apply colors from the rainbow() function. It’s a simple graph, nothing fancy.

Lattice function

I used the bwplot() function from the lattice package to make the next box plot graph.

The syntax is pretty close to the built-in function, which made it easier to work with than I expected. I had to tweak settings to get it to use vertical boxes, and had to find the parameter that would set the colors. Unfortunately I didn’t find an option to draw a line instead of use a plot to mark the median, to match the appearance of the other graphs.

ggplot2 function

For the final graph, I used the ggplot2 package’s ggplot(), geom_boxplot(), labs(), and theme_classic() functions in concert.

It took some time to get the syntax and everything sorted out, since it uses a different syntax from the other two functions. I also had to convert the Time column to factors – without that change, all the percentage data and time data were collected into one giant boxplot. Definitely not what I wanted.

The ggplot2 graph came out looking the best, I think, and next time there would be less fuss given that I’m more familiar with the package.

This module’s assignment was to read data from a file, alter or filter that data, then write the new data to new files.

The code I used is on my GitHub repository here:

https://github.com/jered0/lis4930-rpackage/blob/master/module8/studentsorting.r

I was generally familiar with the concepts involved in reading from files thanks to other coding languages, and the format of read.table is pretty intuitive, so I had no problems there. The file.chooser() function was easy to use, and being able to read the data in all at once without looping through lines one at a time was nice.

It did take me a bit to wrap my head around the ddply function, but running it a few times and tweaking the parameters cleared it up. Being able to fiddle with a data set and pick it apart like that definitely seems like it would come in handy.

I’ve used the grep command on Unix often, so using grepl() was straightforward – though I did have to look over the documentation to see how grep() and grepl() differ, and why grepl needed to be used with subset().