twitter_search.rb

#!/usr/bin/env ruby

require 'rubygems'
require 'twitter'
require 'cgi'
require 'csv'

QUERIES = ['toothache', 'tooth ache', 'dental pain', 'tooth pain']

FIELDS = [
  ['id',         lambda {|s| s}],
  ['from_user',  lambda {|s| s}],
  ['created_at', lambda {|s| DateTime.parse(s).strftime('%F %T')}],
  ['geo',        lambda {|s| s['coordinates'].join(', ') rescue ''}],
  ['text',       lambda {|s| CGI.unescapeHTML(s)}]
]

yesterday = Date.today - 1
results = {}

QUERIES.each do |query|
  puts "Searching for '#{query}' on #{yesterday.strftime('%a, %x')}"

  search = Twitter::Search.new.
    containing(query).
    result_type(:recent).
    since_date(yesterday).
    until_date(yesterday + 1).
    per_page(50)

  page = 0
  search.fetch

  begin
    puts "Page #{page += 1}"
    search.each do |result|
      results[result['id'].to_i] = result
    end
  end while search.fetch_next_page  
end


filename = "#{yesterday.to_s}.csv"

CSV.open(filename, 'wb') do |csv|
  csv << FIELDS.map(&:first)
  results.each do |id, result|
    csv << FIELDS.map {|name, formatter| formatter.call(result[name])}
  end
end

puts "Finished. #{results.size} tweets written to #{filename}"