Files for Productivity Hacks talk.

buffon-needle-configuration.R

# CONFIGURATION ---------------------------------------------------------------

RATIO   = 0.75          # Ratio of needle length to line spacing.
SAMPLES = 500000        # Number of times that the needle is dropped.
SEED    = 13            # Random seed for repeatability.

# -----------------------------------------------------------------------------

buffon-needle.R

# CONFIGURATION ---------------------------------------------------------------

RATIO   = 0.75          # Ratio of needle length to line spacing.
SAMPLES = 500000        # Number of times that the needle is dropped.
SEED    = 13            # Random seed for repeatability.

# -----------------------------------------------------------------------------

buffon <- function() {
  # Sample angle of needle with respect to lines.
  #
  theta = runif(1, 0, pi / 2)
  #
  # Sample the location of the needle's centre. Does the needle cross a line?
  #
  runif(1) <= RATIO * sin(theta)
}

set.seed(SEED)

cross = replicate(SAMPLES, buffon())
#
# Estimate pi.
#
cat(sprintf("ratio = %.3f / %6d samples -> estimate = %.6f\n", RATIO, SAMPLES, 2 * SAMPLES * RATIO / sum(cross)))

configuration-debug-environment.R

DEBUG = as.logical(Sys.getenv("DEBUG", unset = "FALSE"))

print(DEBUG)

configuration-debug-substitution.R

DEBUG = FALSE

print(DEBUG)

configuration-environment.R

# CONFIGURATION ---------------------------------------------------------------

# 1. Get value from environment.
#
DEBUG       = Sys.getenv("DEBUG")
#
# Gives "" if variable is not set.

# 2. Impose a reasonable default.
#
DEBUG       = as.logical(Sys.getenv("DEBUG", unset = "FALSE"))

# SMTP credentials.
#
SMTP_USER   = Sys.getenv("SMTP_USER")
SMTP_PASS   = Sys.getenv("SMTP_PASS")

# -----------------------------------------------------------------------------

configuration-static.R

# CONFIGURATION ---------------------------------------------------------------

N           = 1000      # Number of samples.
K           = 10        # Number of folds.

DEBUG       = FALSE

RECIPIENTS  = c("[email protected]", "[email protected]")

SMTP_SERVER = "smtp.example.com"
SMTP_PORT   = 587
SMTP_USER   = "admin"
SMTP_PASS   = "13*y2YQaV%p5"

# -----------------------------------------------------------------------------

# Rest of script goes here...

if (exists(RECIPIENTS)) {
    if (DEBUG) message("Sending email! ✉")
    # Send email.
}

hello-world-environment.R

# Retrieve "NAME" from environment. Use "World" if not defined.
#
name = Sys.getenv("NAME", unset = "World")

message("Hello ", name, "!")

hello-world-variable.R

name = "World"

message("Hello ", name, "!")

hello-world.R

message("Hello World!")

hello-world.sh

#!/usr/bin/env Rscript

message("Hello World!")

Productivity Hacks - Buffon Needle.md

Some files for talk about Productivity Hacks.

Setup:

Launch a t2.2xlarge instance (8 cores) on AWS.

1. Start a tmux session. Create two horizontal panes (Ctrl-b ").
2. sudo apt update && sudo apt install -y htop r-base
3. wget -O buffon-needle.R http://bit.ly/2HieqWc

run-buffon-needle-multiple.sh

#!/bin/bash

sed 's/\(RATIO   =\).*/\1 0.25/' buffon-needle.R | R --slave
sed 's/\(RATIO   =\).*/\1 0.50/' buffon-needle.R | R --slave
sed 's/\(RATIO   =\).*/\1 0.75/' buffon-needle.R | R --slave
sed 's/\(RATIO   =\).*/\1 1.00/' buffon-needle.R | R --slave
sed 's/\(RATIO   =\).*/\1 1.25/' buffon-needle.R | R --slave
sed 's/\(RATIO   =\).*/\1 1.50/' buffon-needle.R | R --slave

run-buffon-needle-parallel.sh

#!/bin/bash

for ratio in `seq 0 0.125 1`
do
  (sed "s/\(RATIO   = \).*/\1${ratio}/" buffon-needle.R | R --slave &)
done
wait

run-buffon-needle-series.sh

#!/bin/bash

for ratio in `seq 0 0.25 1`
do
  sed "s/\(RATIO   =\).*/\1 ${ratio}/" buffon-needle.R | R --slave
done

run-buffon-needle-single.sh

#!/bin/bash

sed "s/\(RATIO   =\).*/\1 0.5/" buffon-needle.R | R --slave

scattered-configuration.R

DEBUG       = FALSE

#
# Some code goes here...
#

N           = 1000      # Number of samples.
K           = 10        # Number of folds.

#
# Some more code goes here...
#

RECIPIENTS  = c("[email protected]", "[email protected]")

#
# Still more code goes here...
#

SMTP_SERVER = "smtp.example.com"
SMTP_PORT   = 587
SMTP_USER   = "admin"
SMTP_PASS   = "13*y2YQaV%p5"

typical-script-layout.R

# 1. Load libraries 

library(dplyr)
library(tidyr)
library(purrr)

# 2. Load data
#
titanic <- read.csv("titanic.csv")

# 3. Prepare data
#
titanic <- na.omit(titanic)

# 4. Do analysis
#
model <- glm(survived ~ ., data = titanic, family = binomial)

# 5. Disseminate results
#
# - Write to CSV/XLS or database
# - Send email

# 6. Clean up
#
rm(titanic)