Data Science Portfolio

---

Animated Plots with NFL Tracking Data (Shiny app)

Shiny app that generates animations of NFL passing plays from the 2018 season, using a custom-built theme. Inspired by NFL Big Data Bowl 2021 on Kaggle.

Those interested in producing their own American football plots can check-out my ggfootball package in R. To install, simply use the following command: devtools::install_github("robkravec/ggfootball")

---

Atlanta Braves Statcast Data Exploration

Exploration of 2017-2020 Atlanta Braves Statcast data (pitchers only). Sub-projects include:

• Pitch classification using gradient boosted tree model (XGBoost)
• Interactive plots of pitch movement using ggiraph package
• Evolution of Atlanta Braves pitching staff over time, stitched together with patchwork package
• Comparison of actual ball / strike calls to an “objective strike zone”

---

Economic Impact of COVID-19 (Duke Datathon 2020)

Winning submission from Duke 2020 Datathon, which challenged entrants to understand and forecast local-level economic impacts of COVID-19. I specifically worked with unemployment data from the Central Population Survey and highlight this analysis in the linked repository

---

Top News Headlines (Shiny App)

Shiny app that displays top headlines and news sources in tabular form based on user inputs

---

NYC Taxi Visualizations with Spark

Analysis of February 2015 NYC taxi data. Primarily highlights advanced visualization techniques, use of Spark, and reproducible multi-step research processes

---

Blackjack Strategy Simulations with Parallel Processing

Assessment of different blackjack strategies through simulation, emphasizing efficiency through parallel processing and purposeful design of functions