Analysis of Flash Crashes in Economic Data
The purpose of the research project is to analyze a large database of the New York stock exchange data (with a size of 14.5 terabytes) in order to detect financial irregularities. Of particular interest is a “flash crash” a phenomenon in which there is a very deep and volatile fall in stock prices within an extremely short period of time. These crashes can be evidence of quasi-legal trading practices such as quote stuffing or spoofing. The project will focus on developing and implementing a program which can detect these extreme fluctuations in order to determine how common flash crashes are as well as to determine if there are correlations between flash crashes.

My name is Mark and I am a computer science major. I am prepared for the research by the classes I have taken at MIT as well as an UROP research project at the Julia Lab. Through the UROP I got acquainted with the lab and with large data sets. I hope to learn about clever ways to mine a data set. What I find exciting about the project is its flexibility and independence as well as a change to deal with so much data.