Finding the perfect recipe for dinner can be frustrating. Sometimes you don’t have the right ingredients, while other times it’s just hard to find that perfect new dish that happens to use the exact flavors you like.
I was tasked with designing Deliziosa, a machine learning recipe app that makes recipe recommendations based on what a user has in their fridge. It also helps users easily build a shopping list for items that they don’t already have on hand.
UX Design
Info-Architecture
Sketching
Wireframing
Prototyping
User Flows
Sitemap
Wireframes
Brand Guide
Prototype
Figma
Color Contrast
Google Docs
Zoom
Finding recipes that that you will love and that use the ingredients that you have at home can be difficult and time consuming.
People who like to cook, but don’t have time to sift through online recipe sites to find new recipes.
Deliziosa allows you to input several variables and then recommends recipes you will like based on the given information. You begin by inputting your likes, dislikes, and allergies, then proceed to take pictures of the items you want to cook with, and it will give you recipe recommendations. The app uses machine learning to improve your recommendations over time based on recipes you like and save.
I was given only two weeks to design the app from start to finish. The scope of the project was to make sure that users would be able to input the ingredients they have at home, be able to build a grocery list, and then to be recommended recipes.
I was assigned with the task of designing a machine learning recipe app from start to finish, but was given no direction as to what it might look like or how it might function. I was only told that users needed to be able to input the ingredients they have at home, be able to build a grocery list, and then to be recommended recipes.
I was provided access to a preexisting competitive analysis, target audience, and user personas to guide my thinking and designs. Business analysts and team members provided these deliverables to me, and I studied these documents carefully before beginning my design process.
I began the design process by thinking about how the app might function. I knew a main component would be that users will need to be able to input ingredients, but I felt that typing out a list or searching for items from a list would take too long and feel tedious. I then considered both options of building a list by speech and voice. I felt speech could be quick and efficient, but I decided it really wouldn’t save much time compared to taking pictures, and taking pictures would be a more engaging and enjoyable experience. Additionally, I considered the accessibility of each option, and realized a list built by voice would leave out the deaf, mute, and those who just didn’t feel like talking.
I then wrote user stories to begin defining what users would need from the app and what they’d be aiming to accomplish when using it.
I then chose the ones that are most high priority to focus on.
As April, I want to input my groceries so that I can see what recipes I can make based on what I have at home.
As April, I want to input my food preferences, so I can find recipes and cook recipes that support my dietary needs.
As April, I want to add items to a grocery list so I know what to buy to make dinner.
As April, I want to find recipes based on what I have in my fridge.
As April, I want to be able to specify my taste preferences so I can find recipe choices I will like.
As April, I want to save my favorite recipes for later.
I used the user stories to create user flows charts, helping me to define how people would interact with the product. This process also helped me get an idea of what pages would be necessary to design for the MVP.
After defining the features I’d need to build, I began sketching wireframes. This process allowed me to explore all my ideas, that way I could choose the layout that felt the most functional and visually appealing.
From my sketches, I took a less-is-more approach, choosing the designs that would most easily allow users to accomplish their goals and find the information they’d be seeking.
I then began building the wireframes in Figma.
I began adding more detail once the layout looked right, constantly iterating on little things in order to improve the visual appeal.
I spent a large portion of this processing perfecting the machine learning aspect. I wanted to make sure that users would not get annoyed or feel slowed down by inputting their likes, dislikes, and allergies, so I made sure to use leave room for big pictures of tasty food, and easy clicks to highlight preferences.
With the wireframes completed, I began brainstorming the visual branding. I knew I wanted to choose colors that make cooking feel fun and lighthearted. With that in mind, I choose a color palette, font, and icons.
Based on the branding I had defined, I began creating high fidelity mockups. I was careful not to overdo the color; making sure users would not be overwhelmed or distracted when seeking recipes.
I then turned the mockups into a working prototype.
This project was tons of fun to design. I love cooking and
trying new recipes, but in creating this app, I realized I don’t find lots of new recipes. Having an app that recommends new recipes based on your preferences and ingredients makes recipes so much easier to find and makes the whole process of finding what to cook much more exiting and enjoyable.
I learned the power of quality photos. Other projects I’ve worked on have not needed these types of pictures. Finding quality photos made all the difference for this project.
I learned to more deeply value the power of quality feedback. Throughout this process, I checked in often with my mentor and peers to seek feedback and thoughts. By being open to feedback and suggestions, I can create an even more valuable product.
Through the process, I got to explore designing for machine learning. I found machine learning to be a lot of fun to design, and it allowed me to explore lots of ideas and play with what it might look like. I hope that in the future I will have the opportunity to design more products based around machine learning.
