Inspiration
Online shopping often requires users to compare the same product across multiple platforms to find the best combination of price, rating, and delivery speed. We wanted to build a simple tool that brings this comparison into one place and helps users make a faster, more informed decision.
What We Built
ShopIQ AI is a product comparison and recommendation application. Users search for a supported product and ShopIQ compares its price, rating, delivery time, seller, and purchase link across Amazon, Flipkart, and Croma.
The application calculates separate scores for price, rating, and delivery speed and combines them to identify the Best Overall option. It also identifies the Best Price option.
We integrated Google Gemini to generate a natural-language explanation of the recommendation, so users can understand why an option was selected rather than simply receiving a result.
How We Built It
The frontend was developed using React and Vite, while the backend was built using Python and Flask. Product information is stored in a CSV dataset and processed by the backend.
The recommendation system uses an explainable weighted scoring approach:
- Price: 50%
- Rating: 30%
- Delivery speed: 20%
Gemini AI receives the relevant recommendation information and generates an explanation for the user.
We also implemented product-not-found handling so users can see which products are currently supported.
What We Learned
Through this project, we learned how to connect a React frontend with a Flask backend, work with structured product data, design an explainable recommendation system, and integrate the Gemini API securely using environment variables.
We also learned that AI is more useful when combined with deterministic logic. The scoring system makes the recommendation transparent, while Gemini makes the result easier for users to understand.
Challenges We Faced
One challenge was deciding how to compare products using multiple factors rather than relying only on price. We addressed this by creating a weighted scoring system.
Another challenge was integrating Gemini while keeping the API key secure. We solved this by keeping the API key on the backend and using an environment variable.
We also had to handle unsupported searches and make the interface clear enough for users to understand the recommendation and the reasoning behind it.
Future Improvements
Future versions could include real-time prices, additional shopping platforms, price-history tracking, personalized recommendations, review summarization, and real-time inventory information.
Log in or sign up for Devpost to join the conversation.