Analysis of Swertres lotto results.

analysis.R

setwd("~/Dropbox/Projects/suertres-analysis/")

data <- read.csv("data/processed/suertres.csv", header=TRUE,
                 colClasses=c(rep("factor", 6), "integer"))
data$date <- as.Date(data$date, format= "%m/%d/%Y")


length(data)
length(data[,1])


winners.per.date <- tapply(data$num.winners, data$date, sum)
plot(as.Date(names(winners.per.date)), winners.per.date, type="l",
     xlab="Date", ylab="Number of winners",
     main="Number of winners over time")
min(data$date[data$time.slot == "11am"])
min(data$date[data$time.slot ==  "4pm"])
min(data$date[data$time.slot ==  "9pm"])
sum(winners.per.date)
sum(winners.per.date) * 4500


digit1.counts <- tapply(data$digit1, data$digit1, length)
digit2.counts <- tapply(data$digit2, data$digit2, length)
digit3.counts <- tapply(data$digit3, data$digit3, length)
digit.counts  <- t(cbind(digit1.counts, digit2.counts, digit3.counts))
barplot(digit.counts, main="Digit occurences by place",
        xlab="Digit", col=c("darkblue","red", "green"),
        legend = c("1st", "2nd", "3rd"), beside=TRUE)


result.counts <- tapply(data$result, data$result, length)
barplot(result.counts, main="Number of times a winning result has occurred",
        xlab="Winning Result")
min.count <- result.counts[which.min(result.counts)]
max.count <- result.counts[which.max(result.counts)]
result.counts[result.counts == min.count]
result.counts[result.counts == max.count]
mean(result.counts)
median(result.counts)

earliest.occurences <- as.Date(tapply(data$date, data$result, min), origin="1970-01-01")
which.max(earliest.occurences)
max(earliest.occurences)

process.rb

require 'nokogiri'

output_filename = "data/processed/suertres.csv"
input_filenames = {
  "11am" => "data/raw/11am.html",
  "4pm"  => "data/raw/4pm.html",
  "9pm"  => "data/raw/9pm.html",
}

def process(input_filenames, output_filename)
  File.open(output_filename, "w") do |f|
    f.puts("time.slot,result,digit1,digit2,digit3,date,num.winners")
    get_lines_to_write(input_filenames) do |line_to_write|
      f.puts(line_to_write)
    end
  end
end

# this will yield csv strings of the lines we'd write out
# eg: "11am,143,1,4,3,11/16/2006,218"
def get_lines_to_write(input_filenames)
  input_filenames.each do |time_slot, filename|
    row_data(filename) do |data|
      arr = [time_slot] + data
      yield arr.join(",")
    end
  end
end

# this will yield arrays of the data we are interested in
# eg: ['535', '5', '3', '5', '11/8/2006', '155']
def row_data(input_filename)
  html_string = IO.read(input_filename)
  html = Nokogiri::HTML(html_string)

  grid_rows = html.css('table#ctl00_cphContent_gridLotto tr')

  # skip first row since it's a header
  grid_rows[1..-1].each do |row|
    columns = row.css('td')

    result_digits = columns[1].text.strip.split("-")
    date, num_winners = [2, 4].map { |i| columns[i].text.strip }

    yield [result_digits.join] + result_digits + [date, num_winners]
  end
end

process(input_filenames, output_filename)