This Capstone Project is part of my Master’s degree in Data Science at the University of British Columbia, designed to bridge academia and industry by applying classroom knowledge to solve real-world industry problems. Working closely with the Vancouver Whitecaps, I tackled a sports analytics challenge: identifying set-piece (corners and free kicks) strategies correlated with increased scoring chances.

Due to the confidentiality agreement (NDA) signed with the organization, I cannot fully disclose the specific resources, data or methodologies employed. However, my key contributions included:

• Statistical Analysis

  • Analyzed player tracking and event data from the Whitecaps’ Data Warehouse.
  • Identified and validated the features most strongly linked to higher scoring chances from set pieces.

• Predictive Modeling

  • Developed a machine learning model to predict the likelihood of scoring from future set-piece situations under different game conditions.
  • Built a full ML pipeline using object-oriented design patterns and exposed the models through REST APIs with FastAPI, allowing coaching and analytics staff to easily access predictions and apply them in match preparation.

• Exploratory Data Analysis Tool

  • Designed an interactive dashboard for analysts, combining statistical insights and model predictions in a clear, accessible interface.
  • The dashboard acted as the client for our API: analysts entered inputs, the system ran the selected model, and results were displayed visually, making predictive insights usable for non-technical stakeholders.

Tech stack used: Snowflake, Streamlit, FastAPI, SQL, Python