Friday, October 05, 2007

Starting to collect results

So far I've managed to scan every story that relates to my chosen topics. When you search Digg for a story, you get a list of all the stories that were ever submitted to Digg with the search words appearing in the title. First somebody submits a page, and then other people come along and recommend the page. The page winds up with a "score" that indicates the total number of people who recommended the same site.

Collected below is output from my analysis that shows how many stories appeared, followed by the average score of all the stories.

Welcome to News Miner 1.0.
The database has been opened.
What shall we do today, hmmm?
1. Get monthly news clusters (one topic)
2. Get monthly news clusters (all topics)
3. Explore current clusters
4. Get new Digg scores (one topic)
5. Get new Digg scores (all topics)
6. Generate results table
7. Analyze results
0. Quit

There were 14 topics
topic "Anna Nicole Smith": size 752, mean 3.5226063829787235
topic "Barack Obama": size 756, mean 6.994708994708994
topic "Blackwater": size 358, mean 9.997206703910615
topic "Britney Spears": size 3019, mean 3.1570056310036434
topic "Giuliani": size 1468, mean 18.582425068119893
topic "Gulf Coast": size 61, mean 3.459016393442623
topic "Harry Potter": size 2502, mean 5.09912070343725
topic "Hillary Clinton": size 996, mean 7.945783132530121
topic "John Edwards": size 566, mean 7.80565371024735
topic "Mitt Romney": size 426, mean 7.030516431924883
topic "New Orleans": size 858, mean 5.970862470862471
topic "Paris Hilton": size 2497, mean 3.644373247897477
topic "Rupert Murdoch": size 127, mean 7.52755905511811
topic "Tiger Woods": size 390, mean 3.546153846153846

I've highlighted a couple of interesting numbers in bold italics. Notice that "Paris Hilton" and "Britney Spears" both have a very high numbers of pages, indicating that many people found stories about those people that they considered were worth submitting. (Most of them, by the way, are jokes or empty promises of smut.) But other people either don't read those stories, or don't like them enough to recommend them.

On the other hand, look which topic is far and away the clear winner of the Digg scoring game: Rudy Giuliani with an average of 18.6. In fact, Rudy's average score is more than twice as high as the average of his next political competitor Hillary Clinton.

So that means that people love Rudy Giuliani, right? Ermmm... not exactly. Look at the headlines on the page showing the all-time highest rated stories about Giuliani.

  • Mr Giuliani Please Stop Mentioning 9/11
  • Rudy Giuliani Constitutionally Ineligible To Be President
  • Anger at Giuliani 9/11 fundraiser "$9.11 for Rudy" in poor taste
  • America's Worst Nightmare: President Giuliani
  • Giuliani: "For Me Every Day Is An Anniversary Of Sept. 11" GET OFF IT!
  • Rudy Giuliani: "Freedom is Slavery"
  • Rudy Giuliani's daughter is supporting Barack Obama
  • DIGG this! Soldier to Giuliani: Have you done your foreign policy homework?
  • Reporter Arrested on Orders of Giuliani Press Secretary
  • Giuliani Closed Off Streets to Avoid 9/11 Victims' Families

Um, yeah. Are you noticing a pattern here? They're all negative. Not some of them. All of the top nine stories. Apparently people love to read about Giuliani so they can reinforce how much they hate him.

To be fair, Hillary Clinton's page has a lot of negative stories too, but certainly not all of them. Many of the hits are anti right wing coverage, and people apparently reacted well to her TV spoof of "The Sopranos" with Bill.

Disclaimer: Digg users are not a representative sample of the general public, but they are my stand-ins for them. Digg users are a self-selected group of active news readers. What Nixon referred to as "The silent majority," aka "People who don't pay attention" are not represented here, and I have no way of knowing which stories they would recommend if they were asked.

No comments:

Post a Comment