JiayiLu2003/WorldCupROI

What it solves

WorldCupROI is an AI-driven sports sponsorship intelligence platform designed to move beyond simple match prediction. It helps brands determine which sponsorship strategies to choose by analyzing the relationship between match performance, media attention, fan behavior, and sponsor investment, while explicitly accounting for uncertainty and risk.

How it works

The platform uses a multi-layered architecture to convert sports and attention signals into commercial forecasts:

• ML Modeling: It trains match outcome and sponsor ROI models (using baselines like Ridge regression and Centroid classifiers) to predict commercial returns.
• Risk & Explainability: It employs SHAP-style feature contributions for auditability, conformal prediction for reliable intervals, and Monte Carlo simulations to assess risk distributions.
• Graph Intelligence: It maps sponsors, teams, and players as a weighted commercial graph to identify influence patterns and leverage using NetworkX and GCN-style baselines.
• Decision Intelligence: It provides budget-constrained portfolio optimization, causal-effect estimates to separate correlation from causation, and counterfactual stress tests.
• Product Dashboard: A Streamlit-based interface guides users through a workflow of Discover $\rightarrow$ Explain $\rightarrow$ Predict $\rightarrow$ Simulate $\rightarrow$ Recommend.

Who it’s for

• Sports business analysts looking to compare ROI, risk, and sponsor fit.
• ML reviewers who need to inspect model cards, validation, and feature importance.
• Researchers studying the intersection of match performance, text signals, and user attention.

Highlights

• Risk-Aware Decisions: Uses conformal intervals and Monte Carlo risk scores to avoid relying on single point estimates.
• Scenario Simulation: Allows users to test conservative, balanced, and aggressive strategies against a baseline.
• Causal Support: Separates raw correlation from residualized causal-effect evidence for better decision guardrails.
• Graph-Based Discovery: Uses link prediction and node centrality to find future partnership opportunities.