Van Go is a personal art curator that uses natural language processing techniques to analyze descriptions of artworks from museums to help visitors find pieces that match their interests. It tokenizes, lemmatizes, and applies TF-IDF and latent semantic analysis to 100,000 descriptions from the British Museum. This allows it to find artworks with similar themes and styles based on their descriptions. It then demonstrates this by providing recommendations for a visitor interested in landscapes. The recommendations are validated using k-fold cross validation and the Jaccard index to test how well the model performs.