fragrance-finder

Accords-based recommendation system for female fragrances

Table of contents

• General info
• Technologies
• Setup
• Results
• Status

General info

Accords-based recommendation system on 39.7K female fragrances.

This project is part of my fragrance exploration series:

1. K-means++ clustering on fragrance accords
   https://github.com/katarzynajanicka/fragrance-clustering
2. Agglomerative hierarchical clustering on 39.7K female fragrances
   https://github.com/katarzynajanicka/agglomerative-fragrance-clustering
3. Accords-based recommendation system for female fragrances
   https://github.com/katarzynajanicka/fragrance-finder

Technologies

Project is created with Python - version: 3.8.2.

Python libraries:

• fuzzywuzzy - version 0.18.0
• pandas - version 1.1.1
• seaborn - version 0.11.0

Setup

Input data: hierarchical_result.csv, this is the end result of the https://github.com/katarzynajanicka/agglomerative-fragrance-clustering project.

Output data: fragrance_finder.ipynb (Jupyter notebook)

Results

Project structure

Data structure

Most popular fragrances

Findings

Angel Mugler for women

Alien Mugler for women

Coco Mademoiselle Chanel for women

Light Blue Dolce&Gabbana for women

Hypnotic Poison Christian Dior for women

J'adore Christian Dior for women

Flowerbomb Viktor&Rolf for women

Euphoria Calvin Klein for women

Black Orchid Tom Ford for women

Lolita Lempicka Lolita Lempicka for women

La Vie Est Belle Lancome for women

Shalimar Eau de Parfum Guerlain for women

Chanel No 5 Parfum Chanel for women

Twilly d’Hermès Hermès for women

Knot Eau Absolue Bottega Veneta for women

Status

Project is finished.