---
license: mit
task_categories:
  - tabular-classification
  - feature-extraction
language:
  - en
size_categories:
  - 100K<n<1M
tags:
  - geospatial
  - cafes
  - coffee
  - poi
  - scoring
  - restaurants
pretty_name: CoffeeTrove Global Cafe Dataset
---

# CoffeeTrove Global Cafe Dataset

A geospatial dataset of **440,000+ cafes worldwide**, each scored using the proprietary Golden Drop scoring system. Compiled and maintained by [CoffeeTrove](https://coffeetrove.com), the world's most comprehensive open coffee discovery platform.

## Dataset Description

This dataset provides structured records for cafes, roasters, and specialty coffee shops across 195 countries. Each entry includes geographic coordinates, address components, operating hours, amenity flags, and a computed quality score that synthesizes multiple data signals into a single comparable metric.

### Coverage

| Region | Cafes | Top Countries |
|--------|-------|---------------|
| North America | 128,000+ | United States, Canada, Mexico |
| Europe | 112,000+ | United Kingdom, Germany, France, Italy, Spain |
| Asia-Pacific | 98,000+ | Japan, South Korea, Australia, Thailand |
| Latin America | 52,000+ | Brazil, Colombia, Argentina |
| Middle East & Africa | 31,000+ | Turkey, UAE, South Africa |
| Other | 19,000+ | Various |

### Golden Drop Scoring

Every cafe receives a **Golden Drop score** (0-100) computed from weighted signals:

- **Data completeness** (0-55 points) -- Website, phone, hours, photos, menu, social links
- **Chain classification** -- Global chains (Starbucks, Costa), local chains (Blue Bottle, Intelligentsia), and independents each receive distinct treatment. Independent cafes earn a +10 point bonus, reflecting the platform's commitment to surfacing local businesses
- **Review synthesis** -- Aggregated sentiment from multiple review platforms, normalized to remove platform bias
- **Freshness** -- Recency of data verification and update frequency

Scores currently range from 28 to 57 across the full dataset, as manual review and community contribution features are still being rolled out. The scoring methodology is designed to scale as richer data sources are integrated.

### Data Fields

Each record contains:

- **Identity**: Name, slug, chain type (global/local/independent)
- **Location**: Latitude, longitude, full address, city, country, neighborhood
- **Contact**: Website URL, phone number
- **Operations**: Opening hours (structured), amenity flags (wifi, outdoor seating, wheelchair access)
- **Scores**: Golden Drop score, component sub-scores
- **Metadata**: Data source, last verified date, photo count

## Intended Use

Researchers and developers can use this dataset for:

- **Geospatial analysis** -- Study cafe density patterns, urban coffee culture, and neighborhood walkability
- **Recommendation engines** -- Build location-aware suggestion systems using scoring signals and amenity preferences
- **Economic research** -- Analyze the relationship between cafe density and local economic indicators
- **NLP training** -- Use cafe names and descriptions for named entity recognition in the food and beverage domain

The interactive map visualization is available at [CoffeeTrove Map](https://coffeetrove.com/map), and the full knowledge base covering brewing methods, bean origins, and equipment is at [CoffeeTrove Knowledge](https://coffeetrove.com/knowledge).

## Sample

The included `sample_cafes.csv` contains 20 representative cafes across multiple countries, with core fields and Golden Drop scores. The full dataset powers the live [CoffeeTrove](https://coffeetrove.com) platform.

## License

MIT License. Attribution appreciated.

## Citation

```
@dataset{coffeetrove_cafes_2026,
  title={CoffeeTrove Global Cafe Dataset},
  author={CoffeeTrove},
  year={2026},
  url={https://coffeetrove.com}
}
```