Archive

Posts Tagged ‘statistics’

Can Statistics Predict the Body Count in Scream 4?

April 12, 2011 Leave a comment

I’m usually not a big fan of slasher movies, but Scream was an exception. It has a self mocking, ‘we know it’s kinda stupid but it’s entertaining so shut-up!’ attitude which I found very refreshing.

The sequels weren’t as good as the original, but they were ok. And with Scream 4 being released on April 15, the question is: What will the body count be?

Well, you have to follow the rules of the sequel, of course.

There are certain rules that one must abide by in order to create a successful sequel. Number one: the body count is always bigger.

Forrest Wickman and others on Brow Beat went through the data and created a chart detailing the body count as a function of franchise installment number for many popular horror franchises. They were looking to see if the body count “rule” was actually followed. These are the results:

As you can see, in general the body counts do go up with sequel number.

They also came up with an empirical formula for predicting the number of kills in each installment. This takes into account certain factors such as the ‘zombie factor’ and the number of colons in the title of the movie.

Using this formula, they predicted the body count in Scream 4 would be 11.

One can only hope that David Arquette will be one of those 11.

Hurray for using statistics for something useful!

 Update (April 15, 2011) *SPOILER*

Turns out, the actual body count is 15. Meh, not bad.

I Guess I Have to Retire From Blogging Now

April 6, 2011 3 comments

xkcd has (once again) said it all. Basically summarizing my entire blog in a single cartoon.

 

There’s just…nothing left for me to add. What more can be said? What am I going to do with my time now?

I guess I could try knitting…

 

My Favourite Blog Right Now

April 5, 2011 Leave a comment

It is election time in Canada (again), which means a plethora of poll data is just waiting to be analyzed by journalists, pundits and, well me.

So since I like numbers and statistics, my favourite blog for the rest of the month will be TheeHundredEight.com

ThreeHundredEight.com

This site amalgamates poll data and gives a projection of the next parliament. It is also updated every day so you can obsess over stats every single morning!

Now, I know Canadian politics does not have the animosity, craziness, stupidity or entertainment value of American politics, but that is something we Canadians should be proud of; even if its less fun to watch on TV.

While this site may only be interesting to my readers who occupy the intersection of this Venn diagram:

which should be what? Like, 8 people? 12? Meh, whatever. The site also gives poll data by province and even by riding, so even if you don’t like stats you may be interested in seeing which way the wind is blowing in your area.

Since I live in Alberta, and I am a bleeding heart left-wing hippy (which is how conservatives try to paint anyone who is not conservative), my vote will pretty much be useless. But I will still vote for poops and giggles.

Actually, The Globe and Mail also has a nice summary of the party platforms on its website. It is being continually updated as well, so keep an eye on that if you want a nice summary of the different platforms.

Hurray for democracy! And statistics!

How Facebook Connects the World: A Visual

December 14, 2010 1 comment
A Stunning Visualization of Facebook Friend Connections – Click to Enlarge

This beautiful simulation was created by a Facebook intern named Paul Butler. He is an intern on Facebook’s data infrastructure engineering team.

Using a sample of 10 million friend pairs, he wanted to see how those friends were connected geographically. How did he do it?

He explains in a Facebook note, but it is a little bit technical, so I will attempt to translate for the lay reader.

First he started with his 10 million sample pairs. Then he sorted them out by adding up how many friends each city has with each other. For example, how many friends do people in New York have in Chicago? How many friends do people in Los Angeles have in Tokyo? And so on.

So then he tried plotting lines between those cities. Pairs of cities that had more friends between them would have brighter lines. But this turned out to be a bit of a problem. When he tried this,

…a big white blob appeared in the center of the map. Some of the outer edges of the blob vaguely resembled the continents, but it was clear that I had too much data to get interesting results just by drawing lines. I thought that making the lines semi-transparent would do the trick, but I quickly realized that my graphing environment couldn’t handle enough shades of color for it to work the way I wanted.

His graph simply did not have enough shades of blue to make an accurate representation of the numbers. But since he sounds like a very smart guy, he came up with a solution:

I defined weights for each pair of cities as a function of the Euclidean distance between them and the number of friends between them. Then I plotted lines between the pairs by weight, so that pairs of cities with the most friendships between them were drawn on top of the others.

So cities that were closer together got brighter lines. So the line between New York and Chicago is brighter than the line between Los Angeles and Tokyo. Next, he chose to plot lines between cities that have more friendships between them on top of other lines. So the line from New York to Chicago is plotted on top of the line from New York to Scranton.

And Voila! A really nice map of the world drawn entirely with friendships. Paul Butler had some deep thoughts on it:

It’s not just a pretty picture, it’s a reaffirmation of the impact we have in connecting people, even across oceans and borders.

Facebook has gotten some bad press lately. And frankly, I was getting sick of people clogging up my News Feed with Farmville.

But damn, that picture is cool.