Wikipedia IATA railway stations list scrapper, including coordinates. Written on Ruby. Resulting CSV is included.

parse_iata.rb

# Wikipedia IATA railway stations list scrapper, including coordinates
# revision from: 2015-10-20

require 'net/https'
require 'uri'
require 'nokogiri'
require 'csv'

BASE_URL = 'https://en.wikipedia.org'
BASE_PATH = '/wiki/List_of_IATA-indexed_railway_stations'
DISAMBIGUATION_HASH = { '/wiki/Heathrow_station_(disambiguation)' => '/wiki/Heathrow_Central_railway_station' }
IATAS_WITH_DEAD_LINKS = ['XEW', 'XYM', 'YJH', 'YJV', 'ZMI']

def get_html(url)
  sleep(1.0)
  uri = URI.parse(url)
  http = Net::HTTP.new(uri.host, uri.port)
  http.use_ssl = true
  http.verify_mode = OpenSSL::SSL::VERIFY_NONE
  request = Net::HTTP::Get.new(uri.request_uri)

  response = http.request(request)
  raise 'HTTP request failed!' unless response.code == '200'
  response.body
end

def get_coordinates(path)
  html = get_html(BASE_URL + path)

  document = Nokogiri::HTML(html)
  coordinates = document.at_css('span#coordinates span.geo-dms')
  if coordinates.nil?
    coordinates = document.at_css('table.infobox.vcard span.geo-dms')
  end

  latitude = coordinates.at_css('span.latitude').text
  longitude = coordinates.at_css('span.longitude').text

  return latitude, longitude
end

def convert_coordinate(dms_coordinate)
  match = dms_coordinate.match(/(\d+(\.\d+)?)°(\d+(\.\d+)?)′((\d+(\.\d+)?)″)?(N|E|S|W)/)
  coordinate = match[1].to_f + match[3].to_f / 60 + (match[6].nil? ? 0.0 : (match[6].to_f / 3600))
  coordinate *= -1 if 'SW'.include? match[8]
  coordinate
end

def disambiguate_path(path)
  new_path = DISAMBIGUATION_HASH[path]
  return new_path.nil? ? path : new_path
end

def get_railway_iatas()
  html = get_html(BASE_URL + BASE_PATH)

  document = Nokogiri::HTML(html)

  railway_station_codes = document.at('//h2/span[@id="Railway_station_codes"]/..')

  siblings = railway_station_codes.xpath('following-sibling::*')

  CSV.open('railway_iatas.csv', 'wb') do |csv|
    row = []
    row << 'IATA'
    row << 'Name'
    row << 'DMS latitude'
    row << 'DMS longitude'
    row << 'Latitude'
    row << 'Longitude'
    csv << row

#    index = 0
    siblings.each do |sibling|
      break if sibling.name == 'h2'
      next if sibling.name != 'ul'

      items = sibling.xpath('li')
      items.each do |item|
#        index += 1
#        next if index < 130

	text = item.text
        substring_index = text.index(':')
        description = text[(substring_index + 1)..-1].strip
        iata = text[0..(substring_index - 1)].strip

        link_element = item.at('a')
        if link_element.nil?
          latitude = '0°0′0″N'
          longitude = '0°0′0″E'
        else
          path = link_element.at('@href')
          if path.nil? or IATAS_WITH_DEAD_LINKS.include? iata
            latitude = '0°0′0″N'
            longitude = '0°0′0″E'
          else
            latitude, longitude = get_coordinates(disambiguate_path(path.to_s))
          end
        end
        
        row = []
        row << iata
        row << description
        row << latitude
        row << longitude
        row << convert_coordinate(latitude)
        row << convert_coordinate(longitude)
        csv << row
      end
    end
  end
end

get_railway_iatas()

railway_iatas.csv