Spotify's 'Recommended for You' and Google's semantic search both work by finding the nearest neighbors of your query in a high-dimensional embedding space — the same algorithm you're exploring here.
Drag the star to explore how query position changes which documents are retrieved.
Top-3 Results (Euclidean)
python data science
Data Science Cluster
numpy arrays
Data Science Cluster
pandas dataframe
Data Science Cluster
Legend
Euclidean Distance — d(a, b) = √((ax−bx)² + (ay−by)²)
Measures the straight-line distance between two points. Sensitive to the absolute magnitude of coordinates. Good when vectors are normalized to the same scale.
Real RAG systems embed documents and queries into 768+ dimensions (e.g., OpenAI ada-002 uses 1536 dimensions). We show 2D here for visualization — the same nearest-neighbor search happens in high-dimensional space, accelerated by index structures like HNSW or FAISS.