Modeling Critical Locations in the Infrastructure of a City
Dr. Deb Czarneski
Critical locations in infrastructure are roads that if damaged would require a large disruption in the ability of vehicles to navigate the city. For instance a bridge across a major river would be considered critical because removing this bridge would make it more difficult to navigate from one side of the river to the other side. Determining critical locations is important in order to protect a city against a crisis (terrorism, natural disaster, etc.) and to manage repairs after a crisis.
In 2008, Demsar, Spatenkova, and Virrantaus published a paper that determined critical locations in the Helsinki metropolitan area by creating a model and using a computer program to compute the critical locations. This project will recreate and refine the results of these authors using the city of Indianola, Iowa, as our example.
Big League Data
Dr. Bill Schellhorn
The broad topic of this project will be “large data sets”, which we will define as data sets that are difficult to analyze because of their size. Large data sets are being created at an increasing rate due to technological advances in gathering and storing information. Large data sets are common in science, industry, and government.
We will study techniques in statistics and data mining that are key to dealing with large data sets, for instance determining characteristics of the entire data set and detecting trends within the data. The large data sets we will focus on involve sports statistics from Major League Baseball (MLB) and the National Football League (NFL) that are relevant in fantasy sports. Much of the data from MLB will be sabermetric measurements.
In fantasy sports leagues, managers draft actual players to form fantasy teams. Throughout a season, the fantasy teams are matched against each other, with teams awarded wins based on the actual performance of their players. The fantasy league champion is usually determined by a tournament over the final weeks of the regular season. The analysis of sports statistics for fantasy purposes presents many research questions, for example:
- ranking players prior to the draft based on expected performance;
- comparing different drafting strategies;
- designing points systems that encourage competition;
- maximizing the probability of winning a given matchup; and
- predicting team performance via simulations.
Given the abundance of statistical data from MLB and the NFL, investigating these questions will involve using computer software, including spreadsheet applications and the statistical software package R.