ted-postrank.rb

require 'rubygems'
require 'nokogiri'
require 'open-uri'
require 'digest/md5'
require 'fastercsv'
require 'json'
require 'curb'
require 'pp'

data = []
(1..69).each do |page|

  # fetch ted talks from each page
  doc = Nokogiri.parse(open("http://www.ted.com/talks/list/page/#{page}").read)
  talks = doc.search('dd a').inject({}) do |hash,a|
    hash["http://www.ted.com#{a.attributes['href'].value}"] = a.attributes['title'].value
    hash
  end

  # fetch postrank metrics data
  metrics = Curl::Easy.http_post('http://api.postrank.com/v2/entry/metrics?appkey=TEDdemo',
                                 talks.keys.map{|t| "url[]=#{Digest::MD5.hexdigest(t)}"}.join("&"))
  metrics.perform
  metrics = JSON.parse(metrics.body_str)

  talks.keys.each do |url|
    data.push({'title' => talks[url], 'url' => url}.merge(metrics[Digest::MD5.hexdigest(url)]))
  end

  puts "processed page #{page}"
end

# output a CSV file with the results
FasterCSV.open("ted.csv", "w") do |csv|
  columns = data.collect{|d| d.keys}.flatten.uniq.sort
  columns.delete('title')
  columns.delete('url')

  csv << ["Title", "URL", *columns]

  data.each do |a|
    csv << [a['title'], a['url'], *columns.map{|c| a[c] || 0}]
  end
end


# Blog post: http://blog.postrank.com/2010/05/and-the-most-engaging-ted-talk-is/
# Google spreadsheet data: https://spreadsheets0.google.com/ccc?key=tWri7T3f4Ex6-uVU8i9-FFQ&hl=en