PUBLISHED WORK
Featured writer for Towards Data Science
October 2, 2019
WHAT CAN MACHINE LEARNING TELL US ABOUT AMERICA'S GUN LAWS?
Everytown Research maintains a database that tracks changes in state gun laws across 85 features in 8 categories since 1991. We reconcile this data with homicide and suicide data from the CDC to find out how changes in gun laws impact these rates. In addition, we seek to understand how weak laws in one state affect homicide rates in neighboring states. The goal is to find out what practical steps state legislatures can take, if any, to reduce gun violence.
August 9, 2019
THE BUILDING BLOCKS OF BUSINESS STRATEGY
The key to making better business decisions is proper utilization of data. Rather than throwing ideas at the wall to see what sticks, we can draw insights that give business leaders more clarity on what to do and how to do it. This is demonstrated through an example using advanced SQL queries and hypothesis testing with the Northwind Database.
August 6, 2019
LEADING CAUSES OF DEATH IN THE UNITED STATES
The Center for Disease Control’s (CDC) National Center for Health Statistics (NCHS) maintains a database of age-adjusted death rates and counts across the United States for the top 10 causes of death. Grouped by cause, state, and year, the data is available for the years 1999–2016. The objective is to identify the leading cause of death in the United States since 1999, and then identify the states that should be treated as high risk in 2019.
July 31, 2019
USING ARTIFICIAL NEURAL NETWORKS TO ANALYZE PRESIDENTIAL SPEECHES
In most classification tasks, the general aim is to simply maximize some measure of accuracy, whether it’s an F1 Score, Balanced Accuracy, etc. In these cases, we seek to understand the errors for the sole purpose of minimizing their frequency in the future. In general, we want to separate datasets into as clear and distinct groups as possible. But what if we want to do the opposite? What if we have data that is already clearly distinct, but we want to understand how they fit together? In these cases, we can potentially learn more from the errors than we can from the accuracy levels of predictions.