this could be a lot faster by firing off parallel requests with something like http://rubydoc.info/github/typhoeus/typhoeus - though this is for one-off use, so should be okay

all_tracks.rb

require 'net/http'
require 'rexml/document'
require 'CSV'


user = 'benjaminf'
key = '' # public api key
page = 0
total = 1 # set properly by response

CSV.open("#{user}.csv", "wb") do |csv|

  while page < total

    url = "http://ws.audioscrobbler.com/2.0/?method=user.getrecenttracks&user=#{user}&api_key=#{key}&limit=200&page=#{page}"

    xml_data = Net::HTTP.get_response(URI.parse(url)).body

    doc = REXML::Document.new(xml_data)

    doc.elements.each('lfm/recenttracks/track') do |ele|

      csv << [
        user,
        ele.get_text('date'),
        ele.get_text('artist'),
        ele.get_text('album'),
        ele.get_text('name'), # track
        ele.get_text('url')
      ]

    end

    total = doc.elements.each('lfm/recenttracks') {|t| t}.first.attributes['totalPages'].to_i
    puts "loaded page #{page} of #{total}"

    page += 1

  end

end