Inspiration
The inspiration for DishCover came from our love of exploring new cuisines and dishes from around the world. We wanted to create a platform that helps people discover dishes based on their taste preferences, dietary requirements, and cultural interests. Our goal is to make it easier for food enthusiasts to find new dishes they'll love, whether they are cooking at home or looking to try something new at a restaurant.
What it does
DishCover is a food recommendation platform that provides personalized dish suggestions. Users can select dishes they like, and DishCover recommends similar dishes based on flavor profiles, ingredients, cuisine types, and other characteristics. It aims to expand users' culinary horizons by introducing them to new foods that match their tastes.
How we built it
We built DishCover using Python, Streamlit for the user interface, and a machine learning model for food recommendations. We used pandas for data manipulation and Scikit-learn to implement the recommendation algorithm, which leverages cosine similarity to find dishes that are similar to the user's selection. Our dataset consists of various Indian dishes, which we preprocessed to extract relevant features like flavor profiles, ingredients, and course types.
Challenges we ran into
One of the biggest challenges we faced was handling the diversity of ingredients and making sure that the data was clean and properly encoded for our machine learning model. We also encountered issues with managing missing values and dealing with variations in ingredient names, which required us to implement data-cleaning techniques. Integrating the recommendation engine with the user interface in Streamlit while maintaining fast performance was another challenge we overcame.
Accomplishments that we're proud of
We're proud of creating a smooth and intuitive user interface with Streamlit that makes it easy for users to interact with our recommendation system. We're also excited about the accuracy of our recommendation model, which successfully suggests dishes that closely match the user's taste preferences. Our ability to transform a complex dataset into a functional and user-friendly product is something we're particularly proud of.
What we learned
We learned a lot about data preprocessing techniques, feature engineering, and building machine learning models for recommendation systems. We also gained experience in creating interactive web applications using Streamlit. Our biggest takeaway was understanding how to handle complex data and design an algorithm that can make personalized recommendations.
What's next for DishCover
The next step for DishCover is to expand the dataset to include dishes from a wide range of cuisines beyond Indian food, making it a truly global food recommendation platform. We also plan to enhance the recommendation algorithm by incorporating user feedback to make the suggestions even more personalized. Additionally, we aim to develop a mobile app version of DishCover to reach a broader audience and integrate it with restaurant menus, allowing users to discover and order recommended dishes at local eateries.
Log in or sign up for Devpost to join the conversation.