DishFish

Inspiration

In the heart of Atlanta, four friends, all students with unique dietary needs, found themselves struggling to navigate the city's food scene. Faced with the challenge of being vegetarians and having specific dietary restrictions, they embarked on a quest to find suitable restaurants within their budget. The arduous process of scouring menus and hoping for a satisfying meal sparked an idea. Fueled by their shared frustration, these four friends, united by their culinary plight, envisioned a solution. Together, they channeled their determination and expertise to create Dish Fish, an innovative app designed to simplify dining choices, making it stress-free, budget-friendly, and tailored to individual tastes. Their shared experience inspired a revolution in the way people discover and enjoy food.

What it does

On the User Side: Dish Fish is a revolutionary app that transforms the dining experience for users. On the user side, the app allows individuals to input their dietary preferences and budget constraints. Using an intelligent search engine, users can instantly find dishes that align with their needs. For instance, vegetarians like the founders can easily locate meat-free options, while health-conscious users can filter for nutritious meals. The app simplifies the process, ensuring users can discover their perfect dish within seconds, eliminating the frustration of menu scanning, and even uniquely creates combinations of dishes suited to the user. Additionally, Dish Fish empowers users to make healthier choices by tracking calories and providing nutritional data, promoting balanced and nutritious meals.

User Analytics: Dish Fish utilizes advanced analytics to enhance user experiences. By collecting data on users' preferences and dining habits, the app employs adaptive learning. This means that the app learns from users' choices and refines its recommendations over time. These insights provide users with increasingly accurate and personalized dish suggestions, ensuring a delightful dining experience tailored to individual tastes. It also provides information vital to budgeting and dieting with analytics such as how much was spent and calorie count for both.

On the Business Side (Dynamic Pricing Assistance): Dish Fish doesn't just cater to users; it also offers invaluable benefits to restaurants. One of its standout features is dynamic pricing assistance. Restaurants can optimize their pricing based on user preferences and market demand. For example, if a particular dish is in high demand among users, the app provides data to suggest an optimal price, maximizing both customer satisfaction and restaurant profits. This real-time pricing adjustment capability allows restaurants to stay competitive and responsive, especially in a post-COVID-19 landscape where market demands can fluctuate rapidly.

In summary, Dish Fish simplifies the dining experience for users, offering tailored dish recommendations while promoting healthier eating habits. On the business side, it provides data-driven insights, including dynamic pricing assistance, enabling restaurants to optimize their offerings and profitability, creating a win-win situation for both users and businesses.

How we built it

Challenges we ran into

We endeavored to use MongoDB as we saw the numerous advantages present compared to other databases but we struggled as none of us had any experience with MongoDB. We persevered and it made the geospatial location sensing part of our application largely easier and much more coherent allowing for a far better product.

Another problem was creating the NLP model for finding tags out of search queries written in plain English. We finally settled on using the Bert model which had amazing accuracy in categorization solving the problem immensely but it took a lot of trial and error.

There was an additional problem with the search algorithm as there was no precedence before of an algorithm that grouped together meals based on a lot of user preferences that were highly variable. But through market research, we developed our algorithm that gave surprisingly accurate results.

We also had the challenge of finding images for the sample database we created so we used DALLE-3 to create images and will even implement a way for restaurants to make calls to DALLE-3 as well for creating images for restaurants if the restaurant doesn't provide them.

Accomplishments that we're proud of

We made a model for tagging what consumers want in a dish out of plain English and using that data to give affordable dishes that suit consumer needs. The fact we accomplished this model is the accomplishment we are most proud of.

What we learned

We learned a lot about Natural Language Processing and language transformers. We also gained a lot of information about the restaurant industry through market research. We learned about the benefits of using MongoDB and Google Cloud which was invaluable, especially regarding the images. We learned about how to prioritize implementation and adapting to time constraints.

What's next for DishFish

90% of restaurants are small businesses. 77% of consumers will visit a restaurant’s website before dining in or taking out food, according to one survey. This same study found that 30% of diners are turned off by a menu that’s difficult to read, and 30% are turned off by a website that looks outdated. And finally 79% of consumers agree that technology helps improve their restaurant experience. These may seem like random statistics but they all point to one thing, the current model of finding menus is terrible harming businesses and consumers alike. And SMBs in particular generally have the worst-looking menus and websites. Restaurants need a better way to get customers through technological means(as that is what customers want) as a 5% increase in loyalty can lead to up to 95% increase in profits. And dynamic pricing models are at the forefront of increasing margins in a margin-low industry by 200-300%. More and more restaurants are looking for ways to implement dynamic pricing models after COVID-19. Overall there is a market need for a product like this meaning that we will try to improve upon the models and fine-tune the analytics to the point its market ready.