README.md

metadata

license: cc-by-nc-4.0
task_categories:
  - feature-extraction
  - text-classification
language:
  - en
tags:
  - fragrance
  - perfume
  - cosmetics
  - recommendation-system
  - e-commerce
  - retail
  - fragrantica
size_categories:
  - 10<n<100
configs:
  - config_name: fragrances
    data_files: fragrances.csv
    default: true
    sep: '|'
  - config_name: brands
    data_files: brands.csv
    sep: '|'
  - config_name: perfumers
    data_files: perfumers.csv
    sep: '|'
  - config_name: notes
    data_files: notes.csv
    sep: '|'
  - config_name: accords
    data_files: accords.csv
    sep: '|'
dataset_info:
  - config_name: fragrances
    features:
      - name: pid
        dtype: int64
      - name: url
        dtype: string
      - name: brand
        dtype: string
      - name: name
        dtype: string
      - name: year
        dtype: int64
      - name: gender
        dtype: string
      - name: collection
        dtype: string
      - name: main_photo
        dtype: string
      - name: info_card
        dtype: string
      - name: user_photoes
        dtype: string
      - name: video_url
        dtype: string
      - name: accords
        dtype: string
      - name: notes_pyramid
        dtype: string
      - name: perfumers
        dtype: string
      - name: description
        dtype: string
      - name: rating
        dtype: string
      - name: appreciation
        dtype: string
      - name: price_value
        dtype: string
      - name: gender_votes
        dtype: string
      - name: longevity
        dtype: string
      - name: sillage
        dtype: string
      - name: season
        dtype: string
      - name: time_of_day
        dtype: string
      - name: by_designer
        dtype: string
      - name: in_collection
        dtype: string
      - name: reminds_of
        dtype: string
      - name: also_like
        dtype: string
      - name: news_ids
        dtype: string
      - name: reviews_count
        dtype: int64
      - name: pros_cons
        dtype: string
  - config_name: brands
    features:
      - name: id
        dtype: string
      - name: name
        dtype: string
      - name: url
        dtype: string
      - name: logo_url
        dtype: string
      - name: country
        dtype: string
      - name: main_activity
        dtype: string
      - name: website
        dtype: string
      - name: parent_company
        dtype: string
      - name: description
        dtype: string
      - name: brand_count
        dtype: int64
  - config_name: perfumers
    features:
      - name: id
        dtype: string
      - name: name
        dtype: string
      - name: url
        dtype: string
      - name: photo_url
        dtype: string
      - name: status
        dtype: string
      - name: company
        dtype: string
      - name: also_worked
        dtype: string
      - name: education
        dtype: string
      - name: web
        dtype: string
      - name: perfumes_count
        dtype: int64
      - name: biography
        dtype: string
  - config_name: notes
    features:
      - name: id
        dtype: string
      - name: name
        dtype: string
      - name: url
        dtype: string
      - name: latin_name
        dtype: string
      - name: other_names
        dtype: string
      - name: group
        dtype: string
      - name: odor_profile
        dtype: string
      - name: main_icon
        dtype: string
      - name: alt_icons
        dtype: string
      - name: background
        dtype: string
      - name: fragrance_count
        dtype: int64
  - config_name: accords
    features:
      - name: id
        dtype: string
      - name: name
        dtype: string
      - name: bar_color
        dtype: string
      - name: font_color
        dtype: string
      - name: fragrance_count
        dtype: int64

FragDB v3.1 — Fragrantica Fragrance Database (Sample)

The most comprehensive structured fragrance database available. This is a free sample containing 10 fragrances with related brands, perfumers, notes, and accords.

What's New in v3.1

• New field: video_url — YouTube video URLs for fragrances
• Removed field: ownership — no longer included (was empty in source data)
• Total: 67 data fields across 5 files (30 fragrance fields)

Dataset Description

FragDB is a relational database of the fragrance industry containing:

File	Records	Fields	Description
fragrances.csv	10	30	Iconic fragrances with notes, accords, ratings
brands.csv	10	10	Luxury brand profiles
perfumers.csv	10	11	Master perfumer profiles
notes.csv	10	11	Fragrance notes with Latin names, odor profiles
accords.csv	10	5	Scent accords with display colors

Full Database

	Sample	Full Database
Fragrances	10	121,539
Brands	10	7,316
Perfumers	10	2,828
Notes	10	1,825
Accords	10	92
Total Records	50	133,600

Full database available at fragdb.net

Quick Start

Using Hugging Face Datasets

from datasets import load_dataset

# Load all files
fragrances = load_dataset("FragDBnet/fragrance-database", "fragrances")
brands = load_dataset("FragDBnet/fragrance-database", "brands")
perfumers = load_dataset("FragDBnet/fragrance-database", "perfumers")
notes = load_dataset("FragDBnet/fragrance-database", "notes")
accords = load_dataset("FragDBnet/fragrance-database", "accords")

Using Pandas

import pandas as pd

fragrances = pd.read_csv('fragrances.csv', sep='|')
brands = pd.read_csv('brands.csv', sep='|')
perfumers = pd.read_csv('perfumers.csv', sep='|')
notes = pd.read_csv('notes.csv', sep='|')
accords = pd.read_csv('accords.csv', sep='|')

# Join fragrances with brands
fragrances['brand_id'] = fragrances['brand'].str.split(';').str[1]
df = fragrances.merge(brands, left_on='brand_id', right_on='id', suffixes=('', '_brand'))

print(df[['name', 'name_brand', 'country', 'rating']])

Data Structure

fragrances.csv (30 fields)

Identity & Basic Info

Field	Description	Example
pid	Unique fragrance ID	9828
url	Direct link to fragrance page	URL
brand	Brand name and ID reference	Creed;b1
name	Fragrance name	Aventus
year	Release year	2010
gender	Target gender	for men, for women, for women and men
collection	Collection within brand	Text

Media

Field	Description	Format
main_photo	Main product image	URL
info_card	Info card image	URL
user_photoes	User-submitted photos	Semicolon-separated URLs
video_url	YouTube video	URL (new in v3.1)

Composition

Field	Description	Format
accords	Scent accords with strength	a1:100;a2:67;a3:45 (join with accords.csv)
notes_pyramid	Notes by layer	top(name,id,url,opacity,weight;...)middle(...)base(...)
perfumers	Perfumer names and IDs	Erwin Creed;p1;Olivier Creed;p2
description	Fragrance description	HTML text

Ratings & Votes (Structured Format: category:votes:percent)

Field	Description	Format
rating	Average rating & vote count	4.33;24561
appreciation	Love/like/ok/dislike/hate	love:5000:45.2;like:3000:27.1;...
price_value	Price perception votes	way_overpriced:6658:30;overpriced:2844:13;...
gender_votes	Gender suitability votes	female:149:2;unisex:866:10;male:7977:88
longevity	Duration votes	very_weak:784:5;weak:1459:10;moderate:5869:40;...
sillage	Projection votes	intimate:1816:12;moderate:8139:55;strong:4289:29;...
season	Seasonal suitability	winter:4439:44.39;spring:9760:97.60;...
time_of_day	Day/night suitability	day:10000:100;night:6893:68.93

Related Fragrances

Field	Description	Format
by_designer	Same brand fragrances	Semicolon-separated PIDs
in_collection	Same collection fragrances	Semicolon-separated PIDs
reminds_of	Similar fragrances	Semicolon-separated PIDs
also_like	Recommended fragrances	Semicolon-separated PIDs

New in v3.0

Field	Description
reviews_count	Total number of user reviews
pros_cons	AI-generated pros/cons summary with vote counts

notes.csv (11 fields) — NEW in v3.0

Field	Description	Example
id	Unique note identifier	n1
name	Note name	Lavender
url	Fragrantica note page	URL
latin_name	Latin/scientific name	Lavandula angustifolia
other_names	Alternative names	English Lavender, True Lavender
group	Note category	Flowers, Woods, Citrus
odor_profile	Scent description	Fresh, herbal, floral...
main_icon	Primary icon image URL	URL
alt_icons	Alternative icons	Semicolon-separated URLs
background	Background/splash image	URL
fragrance_count	Number of fragrances	12229

accords.csv (5 fields) — NEW in v3.0

Field	Description	Example
id	Unique accord identifier	a1
name	Accord name	woody
bar_color	Display bar color (hex)	#774414
font_color	Text color (hex)	#FFFFFF
fragrance_count	Number of fragrances	45892

brands.csv (10 fields)

Field	Description	Example
id	Unique brand identifier	b1
name	Brand name	Creed
url	Fragrantica brand page	URL
logo_url	Brand logo image	URL
country	Country of origin	France
main_activity	Primary business	Fragrance house
website	Official website	https://www.creed.com
parent_company	Parent company	Kering
description	Brand description	HTML text
brand_count	Number of fragrances	847

perfumers.csv (11 fields)

Field	Description	Example
id	Unique perfumer identifier	p1
name	Perfumer name	Alberto Morillas
url	Fragrantica perfumer page	URL
photo_url	Perfumer photo	URL
status	Professional status	Master Perfumer
company	Current company	Firmenich
also_worked	Previous companies	Quest International, Givaudan
education	Education	ISIPCA
web	Personal website	URL
perfumes_count	Number of fragrances	538
biography	Biography	HTML text

Parsing Examples (v3.0)

Parse v3.0 voting format

def parse_votes(votes_str):
    """Parse v3.0 voting format: category:votes:percent"""
    result = {}
    for item in votes_str.split(';'):
        parts = item.split(':')
        if len(parts) >= 3:
            result[parts[0]] = {
                'votes': int(parts[1]),
                'percent': float(parts[2])
            }
    return result

longevity = parse_votes(row['longevity'])
# {'very_weak': {'votes': 784, 'percent': 5.0}, 'weak': {...}, ...}

Parse v3.0 accords format with join

def parse_accords(accords_str, accords_df):
    """Parse v3.0 accords format: id:percent and join with reference"""
    result = []
    for item in accords_str.split(';'):
        accord_id, percent = item.split(':')
        accord_info = accords_df[accords_df['id'] == accord_id].iloc[0]
        result.append({
            'name': accord_info['name'],
            'percent': int(percent),
            'bar_color': accord_info['bar_color'],
            'font_color': accord_info['font_color']
        })
    return result

Parse notes pyramid with opacity/weight

import re

def parse_notes_pyramid(pyramid_str):
    """Parse v3.0 notes pyramid with opacity and weight"""
    result = {'top': [], 'middle': [], 'base': []}
    for layer in ['top', 'middle', 'base']:
        match = re.search(rf'{layer}\(([^)]+)\)', pyramid_str)
        if match:
            for note in match.group(1).split(';'):
                parts = note.split(',')
                result[layer].append({
                    'name': parts[0],
                    'id': parts[1] if len(parts) > 1 else None,
                    'url': parts[2] if len(parts) > 2 else None,
                    'opacity': float(parts[3]) if len(parts) > 3 else None,
                    'weight': float(parts[4]) if len(parts) > 4 else None
                })
    return result

Use Cases

• Recommendation Systems — Build "if you like X, try Y" engines using accords, notes, and also_like data
• Market Analysis — Analyze trends by brand, country, year, or perfumer
• NLP — Process descriptions, odor profiles, and pros/cons data
• Collection Apps — Build fragrance tracking and discovery apps
• E-commerce — Enrich product catalogs with detailed fragrance data
• Data Visualization — Create accord charts with actual display colors from accords.csv

File Format

• Format: CSV (pipe | delimited)
• Encoding: UTF-8
• Quote Character: " (double quote)

Links

• Full Database: fragdb.net
• GitHub: github.com/FragDB/fragrance-database
• Kaggle: kaggle.com/datasets/eriklindqvist/fragdb-fragrance-database

License

This sample is released under the CC BY-NC 4.0 License. Free for non-commercial use with attribution.

The full database requires a commercial license — see fragdb.net for details.

Citation

@dataset{fragdb2026,
  title={FragDB Fragrantica Fragrance Database},
  author={FragDB},
  year={2026},
  version={3.1},
  url={https://fragdb.net},
  note={Sample dataset with 5 files, 67 fields}
}