I built on a published ranking system to evaluation the CFP selection comittee's current rankings objectively.

I used D3.js and JQuery to build an interactive data visualization with US Census Data.

I used Deep Learning - a mix of pretrained components and trained from scratch components - to build an image captioning system from Microsoft's COCO dataset using Pytorch and deployed it to AWS.

I collected some data and got to the bottom of it.

I used a model trained on tweets from the month prior to the 2014 Congressional Elections to predict the outcome of the 2016 races.

I harvested some tweets from the month prior to the 2014 Congressional Elections and used them to predict the outcome of the races.

Are men paid more than women for the same work? It's a complicated question. I do my best to do this complex question justice.

What are the chances of the world series going 7 games? 6 games? 5 games?

What I think about the current standard of polling and how it could be improved.

Visualize a Bayesian update of a Normal Random Variable.

How well did each of the Ryder Cup competitors perform as individuals? I use Ridge Regularization to help find the answer.

My process of employing unsupervised learning techniques to summarize the topics tweeted about the Rio Olympics.

How have the Salaries of Hall of Famers changed over time? Who is highest paid Hall of Famer?

How well can we predict the opening weekend box office gross from observing tweets for a week before opening night?

On my way to Metis this Friday morning, I noticed the 6 train, which is normally a complete zoo at 8:30 am, was relatively calm. I'd like to investigate whether this was due to the upcoming holiday weeknd, or if this is just a general phenomena on Fridays.

I use Bayes' Theorem to reason about this question.

I use simulation in R to investigate the optimal strategy for placing in the top 5 in this Kaggle competition.

So you're thinking of playing the Prediction Markets? Let me help you build the optimal position according to your beliefs, the price of the market, and your desired level of risk.

I used Python to answer this question in a few different ways. "Frequestist", Bayesian, exact, and simulation.