Makes a tsv of the popularity of all TED talks to date using Postrank, if you have a Postrank API key

ted-postrank.rb

#! /usr/bin/env ruby
# good with ruby 1.8.7
# not good: the Postrank key used accepts no more traffic, and Google has since
# acquired Postrank, so there currently isn't any way of getting a new one. :-(

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

data = []
page = 1
pages = nil # not known until we've read the first page

while !pages || page <= pages
  # fetch ted talk url => title from each page
  doc = Nokogiri.parse(open("http://www.ted.com/talks/list/page/#{page}").read)

  pages = doc.xpath("string(//*[starts-with(.,'Showing page #{page } of ')])")
  pages = pages.match(/\d+$/)[0].to_i
  puts "processing page #{page} of #{pages}"

  talks = doc.search('dd a').inject({}) do |hash,a|
    url = "http://www.ted.com#{a.attributes['href'].value}"
    hash[url] = 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)

  if metrics['error'] then
    puts metrics['error']
    exit 1
  end

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