Fragrance of Data by deepali kank

Exploring the Essence: Insights from Perfume Data Visualization

This project explores a fascinating #TidTuesday dataset on perfumes, sourced from Parfumo, a community of perfume enthusiasts. Web-scraped by Olga G., the dataset covers details of 59,325 perfumes, including ratings, olfactory notes (top, middle, and base), perfumers, and release years. Focusing on top notes—the first impression fragrances leave—the dataset was cleaned using R and visualized through an interactive beeswarm chart created with D3.js.

Size 59,325 perfumes
Details Ratings, olfactory notes, perfumers, release years
Brands 1,442 brands
Perfumes 55,110 unique names
Top Notes 2,430 unique top notes

Using R, the dataset was cleaned and transformed by splitting and reorganizing top notes into individual entries. Key libraries included `tidyverse`, `scales`, `lubridate`, and `janitor`. Missing values were filtered, and note frequencies were calculated.
The data was visualized using an interactive beeswarm chart created with D3.js, showcasing the top notes most frequently used in perfumes from the Parfumo website. Each bubble represents a note, with size indicating frequency. Citrus notes like Bergamot, Mandarin, Grapefruit, and Lemon emerged as dominant, evoking freshness and vibrancy. The design mirrors the elegance of a perfume bottle—bubbles rising as if the fragrance is being sprayed. Gradients and a perfume bottle motif further enhance the luxurious appeal of the visualization.

Analyzing perfume data offered unique insights into the art of fragrance creation. This blend of R and D3.js showcases how visualization can make abstract concepts tangible.