Script for get the visa refusal data in USA by country and by year (2006 to 2015) so far.

usa_visa_refusal_by_country.rb

require 'pdf-reader'
require 'open-uri'
require 'json'

def read_file(year_two_digits)
  url_pattern = "https://travel.state.gov/content/dam/visas/Statistics/Non-Immigrant-Statistics/RefusalRates/FY%s.pdf"
  file = open(url_pattern % year_two_digits)
  PDF::Reader.new(file)
end

def format_year(year)
  year.to_s.rjust(2, '0')
end

def get_data(page)
  page.split("\n").map(&:strip).map { |e| e.gsub(/\s+/, ' ') }.select { |e| e.match(/^.*%$/im) }.map { |e| e.split(/ (?=\d+)/) }
end

import_scope = (6..15).to_a.map(&method(:format_year))

data = {}

import_scope.each do |year|
  pdf = read_file(year)

  countries_with_refusal_rate = []

  pdf.pages.each do |page|
    countries_with_refusal_rate << get_data(page.text)
  end

  data[year] = countries_with_refusal_rate.flatten(1)
end

puts data.to_json