Generally, busy couples eat repetitively and tend to use the same palette of meals every week. To overcome the challenge of selecting a new meal, we present a ticket-printing device. Based on your eating habits, user profile, wind-chill temperature, and day of the week a recipe is selected using Machine Learning (k-nearest-neighbors and content-based filtering algorithms). The recipe is printed on a ticket, containing ingredients and basic cooking instructions, and can be torn off when leaving the house. If a user dislikes the recipe, a new recipe will be printed automatically. The grocery purchase data is used as feedback to the system, improving its ability to select the optimal recipe.
We initiated this design process with extensive user research. Using triangulation, we obtained insights into the habits of our selected target group. Several methods were used, including cultural probes (a personalized journal and website) and contextual inquiries. We generated various concepts based on the insights, which were evaluated with experts with a focus on different interaction types. The selected concept has been further developed to a high fidelity prototype made of multiple materials (concrete, veneer, internal MDF) using various rapid prototyping techniques (laser cutting and vacuum forming).