The Framework

The Flavor Anatomy

How we deconstruct complex spirits into measurable data points.

1. The Flavor Radar (Sensory Identity)

The Flavor Radar acts as the "fingerprint" of a spirit. By normalizing thousands of tasting notes through our NLP engine, we map the relative intensity of core flavor groups. This visualization allows for an immediate comparison between an individual bottle and the industry standard.

Learn more about our methodology in Flavor Profiles.


2. Character Profile (Structural Balance)

The Character Profile translates subjective community sentiment into structural data. Unlike flavor, which describes "what" you taste, character describes "how" the spirit behaves.

Understanding the Metrics

Our engine categorizes mentions on a 5-point scale based on the volume of specific descriptors:

Scores 1-2 (Left): "simple," "flat," or "basic." Scores 4-5 (Right): "complex," "layered," or "premium."
Taste & Complexity

Identifies the sensory depth. This metric distinguishes between a straightforward, consistent pleasure and an evolving, multi-layered journey.

Value & Presentation

The price-to-quality ratio. It reveals if the brand's aesthetic intent and marketing align with the actual liquid quality.

Utility (The Usage Anchor)

Practical Application: Does the community treat this as a versatile "Mixer" for cocktails or a dedicated "Neat / Sipping" spirit?

Among Its Peers (the Peer Map)

The Peer Map on a drink page places the drink among the other drinks of its category, on two of these scales: Simple to Complex (left to right) and Mixer to Neat / Sipping (bottom to top). The star is the drink you are looking at. If you want an approachable sipper, look top left; for a complex bottle to sip neat, top right.

  • The position on each scale counts only the reviews that mention it: the share that says complex minus the share that says simple, from −1 to +1. A review that says nothing about complexity (score 3) does not pull the drink to the middle.
  • Only with enough data: a drink appears once at least five reviews mention each scale. A drink below that is left out and its own page shows no map, and neither does a page whose category has fewer than three drinks on the map.
  • A Spirit Drink (a spiced rum or a flavoured whisky, for example) is only compared with Spirit Drinks of the same base spirit: Rum-Based with Rum-Based, Whisky-Based with Whisky-Based. The category average on its flavor radar follows the same rule. A Spirit Drink whose base spirit is unknown has no map and no average.
  • Hover over a point to see the drink and its number of reviews; click it to open that drink.

3. Strategic Sensory Analysis

By combining these two visualizations, we provide a holistic view. A bottle with a "Floral" radar and a "Heavy Body" character offers a fundamentally different experience than a "Floral" spirit with a "Light Body." This multi-layered approach is the foundation of our precision matching engine.


4. Under the Hood

The models and software behind this page, and when they run.

  • fast-langdetect Google Cloud Translation

    Language detection & translation: A review in another language is translated to English first, because the analysis models read English. The original text stays on the site.

    When: Every new review, written here or collected from other sites.

  • spaCy en_core_web_lg

    Flavor recognition: Finds the terms of our Flavor Dictionary in a review: exact matches, plus close relatives through 300-dimensional word vectors (a cosine similarity of at least 0.75).

    When: Every review.

  • spaCy en_core_web_lg

    Character & occasions: Word lists read mouthfeel and structure (such as smooth or harsh) and the occasions people mention from the text. Community members set these themselves in the review form.

    When: Every review collected from another site.


Deep Dives & Expert Resources

Flavor Frameworks
The Guide
How It Works
Sensory Science
Drinks Data