Explore the Data

We use the Kaggle Wine Reviews dataset to explore whether we can use textual descriptions of wines to predict price and quality. Here's some of the initial exploratory analysis we performed on our dataset.


Points Distribution

Points Distribution

Price Distribution

Price Distribution

Points by Country

Points by Country

Median Price by Country

Price by Country

Mean Price by Variety

Price by Variety

Textual Features

Word Cloud

Topic Modeling using Latent Dirichlet allocation

  1. wine tannins fruit years rich structure wood age ripe firm
  2. finishes easy pretty alcohol drinking pie taste sweet like lot
  3. great wine fruit vineyard freshness wines vintage shows elegance grapes
  4. blackberry cabernet black chocolate blend tannins flavors sauvignon cherry merlot
  5. flavors finish palate aromas fruit bit nose berry plum herbal
  6. citrus apple flavors finish aromas wine white fruit fresh green
  7. aromas spice cherry wine red berry fruit notes palate black
  8. wine acidity red mouthfeel fresh ripe fruit soft fruity flavors
  9. flavors wine imported acidity vanilla sweet oak chardonnay rich style
  10. flavors dry cherry cherries pinot little wine oak good cola