P( ssp3nc3r | Columbian )
generative modeling. bayesian inference. writing. data visualization. sports.
NYC bike share rebalancing information graphic
We continue exploring the NYC bike share program. This time we create an information graphic visualizing bike and docking station availability by time of day and rate of new rides per minute.
Approximating the components of Lupi's Nobel no degrees
Giorgia Lupi is a master at designing information for visual narrative — telling stories through visualization. We study her design and approximate components with code in R.
Exploring CitiBike ride data
NYC Open Data have interesting data including CitiBike ride information. Let’s begin wtih a basic exploratory visual display.
Scoping a data analytics project
Scopes of data analytics projects include various considerations. We provide an overview.
Baseball Abstract 01
Exploratory data analyses’ fun side includes abstract art.
Estimating the marginal value of baseball events
Teams score — or
— event by event. Events change the state of the game, and whether a score results from an event depends upon prior events. By the events a player generates, or more accurately, by the change in expected runs to which they contribute, we can measure players’ offensive values.
McElreath's Statistical Rethinking examples recoded into Stan
McElreath’s Statistical Rethinking is among the best books for learning Bayesian statistics. As an exercise in pedagogy, I’ve recoded all his examples directly into Stan, a probabilistic programming language.
Jointly modeling choice to swing with ball contact
Batters do not usually swing at balls without believing they can make contact. As the two events lack indepencence, we explore a joint model.
Stan Is Not an Acronym: Bayesian Solutions to Sabermetric Shortcomings
Description of our presentation at Saberseminar discussing Bayesian solutions to sabermetric shortcomings.
Visual exploration of starting pitcher duration
Managers do not generally pull starting pitchers to avoid the debated “times through the order penalty,” as shown using a simple survival model. A sketch of other factors are briefly explored.