Nurse advocates for improved disaster planning

She has seen houses ripped from their foundations, elderly people evacuated from their homes without their medications and rescue dogs whose feet were burned in the rubble of the World Trade Center.

## March Madness: Are statistical models more accurate than guessing?

You may have heard the name Nate Silver during the recent presidential election. The New York Times' polling blogger correctly predicted the winner of all 50 states and the District of Columbia. His secret? Statistics.

Silver first caught the public's attention with the creation of a computer system used for evaluating baseball player's stats, which outperformed the analysis of many experts. He then went on to become a full-time online poker player, before turning to politics. During the 2008 presidential election, Silver correctly predicted the winner of 49 of the 50 state using statistical modeling. Due to the accuracy of his forecasts, his blog, FiveThirtyEight, has become a source for data-driven analysis of current topics (sports, politics, and culture).

Since predictive modeling seemed to work well in other arenas- why not create one for March Madness? In 2011 Silver created a NCAA Men's Basketball Forecasting model that used a few simple tenets to determine success- past performance, current performance, coupled with more advanced calculations. For basketball fans (and geeks like us) his model is a thing of wonder. You can access the simple and advanced explanation here. He threw out some of the minutiae in the model because he found that it didn't make a difference (e.g. coach's experience, free throw shooting). He seems to account for many of the "what ifs", but still encourages aficionados to do what a model may not be able to do- think strategically- how a player may match up against one another on a particular day in a particular game.

Here at SAASI, we decided to put Nate Silver's mathematical model to the test! And since it's March Madness, what better way to investigate its accuracy than by comparing brackets?Earlier this week four of our staff members filled out their brackets. From picking favorite mascots and colors to guessing at random, each used a different approach to predict which teams would advance in the NCAA Men's Basketball Tournament. Throughout the rest of the month we'll be comparing their brackets to Silvers' (which was published on his blog).

How they chose...

• Chris: Chose Indiana because that is where Larry Bird is from and he was a great Boston Celtic.
• Davis: Follows basketball and knows what he is doing.
• Emily: Colors and mascots.
• Zoraya: Random picks.

04/09/2013 UPDATE
Final rankings showed that Nate Silver's bracket consistently remained in the first or second place throughout the whole tournament, with the gap between his bracket and our brackets widening as the tournament progressed.

1st place: Nate Silver
2nd place: Davis Brigman
3rd place: Emily Johnson
4th place: Chris Knickerbocker
5th place: Zoraya Cruz-Bonilla

04/02/2013 UPDATE
Currently, rankings remain unchanged.  There are a few more games left to play, but Nate Silver is emerging as the clear victor in this bracket contest.

04/01/2013 UPDATE
After the Regional Semi-finals, Nate Silver is first place! Emily Johnson and David Brigman are tied for second place. Chris Knickerbocker is third place. Zoraya Cruz-Bonilla are last place.

After the 3rd round, Emily Johnson, Nate Silver, and Davis Brigman are all tied! Chris Knickerbocker is second place.  Zoraya Cruz-Bonilla is third place.

After the 2nd round, Emily Johnson are first place! Nate Silver and David Brigman are tied in second place. Chris Knickerbocker is third place.  Zoraya Cruz-Bonilla is last place.

03/25/2013 UPDATE
Davis Brigman and Emily Johnson are tied for first place! Zoraya Cruz-Bonilla is now last place.

03/22/2013 UPDATE
Silver is tied for last place with Chris Knickerbocker. Zoraya Cruz-Bonilla, Emily Johnson, and Davis Brigman are tied for first place!

## The Case of the Forgotten Something or Other

The following is a Prezi presentation using storytelling.

As data detectives, we often use our manifying glass to find the story. In fact, just when we think that there isn't one, we stumble upon another clue that often leads us on a tasty trail.

Click the center to join on the fun.

## Go for the Gold: Using Economics to Make Predictions

Like many Americans this summer, you probably tuned into the 2012 Olympic  Games to watch your favorite sports and favorite athletes go for the gold. While you were rooting for Michael Phelps and Ryan Lochte, did you try to guess how many medals the United States would come home with this year? Whether they would come out on top against close competitor, China? Guess no more; an economics professor has developed a model to predict Olympic medal wins with 94% accuracy.

Dan Johnson is an economics professor at Colorado College. He has been predicting medal winners since the 2000 Summer Games in Sydney using his model. It started off as an experiment to see if economics really could be used to predict medals. Johnson didn't actually think it would work, but it did. The incredible thing about his predictions though is that he doesn't use data from individual athletes or their events. He uses strictly economic measures like the country's per-capita income, population, political structure, climate, and the home-field advantage if they are hosting the games.

This year he correctly guessed that the United States would win the most medals with 99 total medals, while the actual count was 104. He also predicted that Russia would come in second and China third; in reality the two were switched. However, his prediction of Great Britain coming in fourth and Germany fifth proved to be right on the money as well as his prediction of Russia winning 82 medals.

So if models can be used to predict something that has so many variables and emotions like a sporting event, what else can they be used to predict? This may come as a surprise, but economic models are used all the time here on the Binghamton University campus. Enrollment Management, for example, uses models to make predictions about incoming students. One of their main statistics is yield rate (percent of accepted students who actually enroll). Enrollment Management uses yield rates and other variables to determine how many students to accept each year. Residential Life uses similar models to determine the number and type (male/female) of housing vacancies there will be on the first day of the term. Like Professor Dan Johnson does with his model, these models are constantly tweaked using new data to improve their accuracy.

To leave you with some food for thought... how does your department use statistics and models? They could be used to predict anything from how many office supplies to order to how many students to accept each year. So keep track of your statistics and use those numbers to your advantage! In the long run it could save you time, money, stress, and help you make well-informed decisions.