I come with expertise in regression, classification, clustering, modeling and visualization. My capabilities continue to expand as I learn more in my business analytics master’s program as well as on my own.

Multiple newsrooms have called upon me to analyze site and social media analytics, tease out what drives success and develop ways to help reporters and editors incorporate those qualities into more content.


For my capstone project in my Business Analytics master’s program, I used a gradient boosting machine to calculate the likelihood that a publicly traded company would face shareholder class-action litigation and calculated how much directors and officers insurance it should buy. The project involved selecting the most relevant features from a dataset offering about 1,000 possible options, modelling the company’s litigation likelihood, contextualizing that figure within the company’s industry, finding the potential worst-case scenario, and putting together a report to present all my findings.

Another recent project delved into using predictive analytics to project an college athletic departments’ spending figures. That question was answered with a linear regression model. I also trained models in logistic regression, linear and quadratic discriminant analysis, k-nearest neighbors, decision trees, bagging, boosting and random forests in an attempt to see how well the models could categorize which athletic programs belonged to the NCAA’s “Power Five,” “Group of Five,” or “other.” You can take a look at that work in PDF format or on Github.

By creating regression and clustering models for a text-mining dataset provided by UC Irvine, I analyzed what parts of a reporter’s writing style can push a Mashable article closer to viral success, and to what degree they do so.


I’ve worked with in-house visualization tools in previous newsrooms (unfortunately, those projects were lost in a newsroom CMS change). My best project was put together the day of the Supreme Court gay marriage ruling. Using The Dallas Morning News’ Chartwerk tool, I led a team of a dozen reporters calling courthouses in Texas’ 254 counties to find out how they were reacting to the ruling. Each county was color coded based on four classifications of their approach to handling gay marriages at the time. Then when the state’s attorney general issued an opinion on how to handle same-sex marriages over the weekend, we went right back to the phones on Monday morning.

My data visualization journey more recently has involved dives into R Shiny on multiple occasions and Tableau.


As an avid college basketball fan, I have spent much of March Madness toying around with team and player data. It’s an opportunity to both improve my proficiency with R and make complex sports analytics friendly enough for a casual audience. Here’s a look at some of that work: