How do I find the bottleneck?

Writing Efficient R Code

Colin Gillespie

Jumping Rivers & Newcastle University

Code profiling

The general idea is to:

Run the code
Every few milliseconds, record what is being currently executed
Rprof()comes with R and does exactly this
- Tricky to use
Use profvis instead

IMDB data set

From the ggplot2movies package

data(movies, package = "ggplot2movies")
dim(movies)

58788    24

Data frame: around 60,000 rows and 24 columns
Each row corresponds to a particular movie

Braveheart

braveheart = movies[7288,]

Year	Length	Rating
1995	177	8.3

Example: Braveheart

# Load data
data(movies,
     package = "ggplot2movies")
braveheart <- movies[7288,]
movies <- movies[movies$Action==1,]

plot(movies$year, movies$rating,
     xlab = "Year", ylab = "Rating")

# local regression line
model <- loess(rating ~ year,
               data = movies)
j <- order(movies$year)
lines(movies$year[j],
      model$fitted[j],
      col = "forestgreen")

points(braveheart$year,
       braveheart$rating,
       pch = 21,
       bg = "steelblue")

Profvis

RStudio has integrated support for profiling with profvis
- Highlight the code you want to profile
- Profile -> Profile Selected lines

Command line

library("profvis")

profvis({

data(movies, package = "ggplot2movies") # Load data
braveheart <- movies[7288,]
movies <- movies[movies$Action == 1,]
plot(movies$year, movies$rating, xlab = "Year", ylab="Rating")
model <- loess(rating ~ year, data = movies) # loess regression line
j <- order(movies$year)
lines(movies$year[j], model$fitted[j], col="forestgreen", lwd=2)
points(braveheart$year, braveheart$rating,
     pch = 21, bg = "steelblue", cex = 3)

})

Which line do you think will be the slowest?

Let's practice!

Writing Efficient R Code