Estimating physical properties of aromas and fragrances

The challenge

A chemical company markets perfumes and fragrances made up of raw materials and components, each with physical properties that affect transport. For example, a component with a low flash point requires ensuring a cold chain.

Objectives

To develop a solution integrated into the production environment that allows for the analytical estimation of physical properties (flash point, refractive index, and density) of perfumes and fragrances based on the physical properties of their components.

How have we done it?

Machine Learning
MLOps
Django

Vue.js
Securization

Results

The solution accurately estimates physical properties (errors of 5%) and is integrated into the client’s production environment with a web interface.