End to End experiments Personalised Beverages Design
Datasets usually provide raw data for analysis. This raw data often comes in spreadsheet form, but can be any collection of data, on which analysis can be performed.
The food industry aims to create and supply food and drink products that satisfy their customers. Currently, products cannot be tailored to meet each individual's desires, only to the collective preferences of target markets. Is it possible to achieve the former, creating food products that match the individual consumer's explicit preferences? To achieve this, we must discover how consumer's perceive products. Once their perceptions are understood, we can tailor ingredient quantities to match the individual's desired product. We will test this concept on orange beverages.
This work aims to test if we can create tailored orange drinks from a consumer's description of their ideal drink. Their description is converted into ingredient quantities, with which a drink is created using our beverage formulator. The consumer then tastes the drink and assesses if it correctly matches what they requested. This will be experiment 2. To achieve this, we must first learn how people's perceptions match to ingredients. Consumer's will taste orange drinks in which ingredients quantities differ and we ask them to describe how intense they perceive its flavour attributes. This will be experiment 1.
• In experiment 1, we learn how intense consumers perceive flavour attributes of orange drinks for a given set of ingredients quantities.
• In experiment 2, we test if we can choose quantities of ingredients that correctly match consumer's preferences for the same flavour attributes.