World Cup Graph AI Game
Explore hidden World Cup player relationships and turn squad cohesion into a game.
This app follows the attached brief: graph-powered trivia, match retrodiction, cultural edge discovery, and a data pipeline that fuses MongoDB, Neo4j, Qdrant, and OpenAI-backed inference.
Platform pillars
- MongoDB for players, edges, matches, and sessions
- Neo4j sidecar for graph-native traversal
- Qdrant for semantic retrieval and embeddings
- OpenAI for chat, extraction, and inference workflows
Game modes
Each mode maps directly to the brief's core loops and APIs.
Six Degrees of the World Cup
Connect two World Cup players through the shortest chain of graph edges.
- • Shortest-path search with obscure-edge bonus scoring.
- • Time pressure on every hop.
- • Ideal first mode for validating graph connectivity.
Retrodict the Match
Predict historical World Cup outcomes before revealing the result.
- • Compare user intuition with the model's squad graph score.
- • Supports future GNN-backed probabilities.
- • Best fit for match and squad APIs.
Squad Builder
Draft a World Cup XI that maximizes hidden chemistry and graph density.
- • Assemble a lineup and inspect connection quality across positions.
- • Designed to surface non-obvious teammate and culture edges.
- • Great mode for proving graph richness in live demos.
Edge Detective
Identify the hidden attribute linking two players.
- • Multiple-choice edge guessing under time pressure.
- • Great for showcasing obscure graph facts.
- • Directly benefits from LLM-extracted cultural edges.