Description
The main point of this lab is to gain experience with measuring execution time and trying to
write faster code. We will work with a simple CSV file and calculate test error for a simple
linear regression model, exploring the time required for several different approaches.
The Data
The data set for this lab is a single large CSV file containing electric vehicle trips (same as for the
previous lab). This file is available on the VMs at the following location:
/course/data.austintexas.gov/Dockless_Vehicle_Trips.csv
The Task
1. Read the first 100,000 rows from the CSV file using three different approaches:
• read.csv() without specifying colClasses.
• read.csv() specifying colClasses.
• data.table::fread() specifying colClasses.
In each case, skip the first line of data (because it just contains missing values).
Compare the execution time for all three approaches. Check that all three approaches
produce the same set of data values.
2. Read the entire CSV file into R with data.table::fread() and assign the result to a symbol
called tripDT.
3. Measure the time taken to run the following R code.
tripDTsub <- subset(tripDT, Trip.Distance > 0 & Trip.Duration > 0)
tripDTsub$logDuration <- log(tripDTsub$Trip.Duration)
tripDTsub$logDistance <- log(tripDTsub$Trip.Distance)
4. Try to perform the same subsetting and transformation much faster by writing R code
that uses the special data.table syntax. NOTE: your code should create new variables in
tripDT rather than creating a new tripDTsub object.
Check that your new R code produces the same set of data values as the original code (for
the subset of rows with positive distance and duration values).
5. Generate test/training group labels by running the following code.
labels <- rep(1:10, length.out=nrow(tripDT))
groups <- sample(labels)
Use the function mse() (shown below) to estimate the test error for a simple linear regression
model using k-fold cross-validation.
mse <- function(i, formula) {
## cat(i, “\n”)
testSet <- groups == i
trainSet <- groups != i
1
fit <- lm(formula,
data.frame(x=tripDT[trainSet, logDistance],
y=tripDT[trainSet, logDuration]))
pred <- predict(fit, data.frame(x=tripDT$logDistance[testSet]))
mean((pred – tripDT$logDuration[testSet])^2, na.rm=TRUE)
}
Measure the time taken to calculate the test error and profile the calculation to see where
most of the time was spent.
6. Based on your profiling results, try to write new R code to calculate the test error that runs
faster. HINT: one way to speed up the code involves the na.action argument to the lm()
function.
Check that your new R code generates approximately the same answer as the original code.
Please seek assistance from your lecturer if you get stuck with this task – it is not trivial.
7. Use the shell command time to measure how long it takes to process your R Markdown
document.
You will lose marks if “user” plus “system” is longer than 2 minutes.
The Report
Your submission should consist of a tar ball (as generated by the tar shell tool) that contains an
R Markdown document and a Makefile and a processed version of your R Markdown document,
submitted via Canvas.
You should write your document and your Makefile so that the tar ball can be extracted into a directory anywhere on one of the virtual machines provided for this course (sc-cer00014-04.its.auckland.ac.nz
or sc-cer00014-05.its.auckland.ac.nz) and the markdown document can be processed just
by typing make.
Your report should include:
• A description of the data format.
• A description of each of “The Task”s.
• A conclusion summarising the analysis.
Your report should NOT be longer than 10 pages.
Marks will be lost for:
• Submission is not a tar ball.
• More than 10 pages in the report.
• R Markdown file does not run.
• R Markdown file takes longer than 2 minutes to run.
• A “Task” is missing.
• R Markdown file is missing.
• Processed file (pdf or html) is missing.
• Makefile is missing.
• Significantly poor R (or other) code.
2