VeggieWise AI

Inspiration

Small vegetable vendors often struggle to decide how much stock to buy each morning. Overstocking leads to food waste, while understocking results in lost sales. We wanted to help vendors make smarter decisions using AI and data analytics.

What it does

VeggieWise AI predicts vegetable demand and recommends daily purchasing quantities using historical sales, mandi prices, weather forecasts, and festival information. It aims to reduce waste, avoid stock shortages, and improve vendors' purchasing decisions. Kannada voice input and WhatsApp recommendations are planned features.

How we built it

Our planned solution uses Python for data processing, Prophet or LightGBM for demand forecasting, and FastAPI for backend services. Market prices, weather information, and historical sales data will support predictions. A simple dashboard will display suggested quantities and explanations.

Challenges we ran into

Key challenges include obtaining reliable datasets, handling changing demand, combining different data sources, and generating useful predictions for vendors with limited technical knowledge.

Accomplishments that we're proud of

We developed an idea focused on a real-world problem affecting small vendors. Our goal is to make data-driven decision-making accessible, practical, and easy to understand.

What we learned

We explored how data engineering, machine learning, and real-time analytics can be combined to solve practical problems. We also learned the importance of explainable recommendations and accessible interfaces.

What's next for VeggieWise AI

We plan to improve forecasting accuracy, integrate live market and weather data, add Kannada voice-based sales entry, and deliver personalized daily recommendations through WhatsApp

Built With

• apachekafka
• apachespark
• artificialintelligence
• dataengineering
• datascience
• fastapi
• javascript
• machine-learning
• predictiveanalytics
• python