Data methodology

How we use the wine dataset

In short

When our guides mention a typical price, a common region, or a critic-score range, those numbers are aggregate statistics computed from a large wine-review dataset. We never reproduce the original review text.

What we compute

We analyze a public dataset of roughly 150,000 wine reviews. From it we calculate only aggregates — counts, median and typical price ranges, critic-score ranges, and which regions and grapes appear together most often.

  • Typical price: the median and the middle-50% range — computed only from the entries that list a price (some rows have none).
  • Critic scores: the minimum, median, and maximum on a 100-point scale.
  • Common regions and grapes: the most frequent, by number of wines.

What we never do

We do not copy, quote, or paraphrase the individual review descriptions in the dataset. Those belong to their authors. Every sentence you read on our guides is written originally for this site and reviewed by a person for accuracy.

Caveats worth knowing

Any dataset reflects its own coverage. This one skews toward wines reviewed in the United States and is a historical snapshot, so we treat the figures as directional context — a sense of "typically around this price" — rather than a live market quote. We refresh the aggregates when the underlying data changes.