Inspiration
The inspiration for Twala has evolved from a simple marketplace to a powerful, AI-driven assistant for African farmers. While our initial goal was to connect farmers directly with buyers, we recognized that a more fundamental challenge exists: empowering farmers with the data and knowledge they need to optimize their farming practices. Many small-scale farmers lack access to critical information on weather, crop diseases, and market prices, leading to unpredictable yields and low profitability. Our pivot was inspired by the need to put cutting-edge technology directly into the hands of these farmers, providing them with actionable insights to achieve sustainable agriculture and improve their livelihoods.
What it does
Twala is now a comprehensive smart-farming platform that uses AI to provide personalized agricultural advice. The web application features several key modules:
- Disease Detection: Farmers can upload an image of a crop, and our AI model will instantly identify potential diseases and offer treatment recommendations.
- Weather Forecast: By simply entering their location, farmers get real-time weather updates and a 7-day forecast to help them make informed decisions about planting, irrigation, and harvesting.
- Crop Advisory: Based on user-selected crop types and real-time weather data, the platform provides tailored advice on what to plant and when.
- Market Prices: The app tracks current market prices for various crops in different regions, helping farmers plan their selling strategy to maximize profits.
- Soil Parameter Monitoring: For farmers with access to basic IoT sensors, the platform can display real-time data on soil pH, moisture, temperature, and nutrients, offering a crucial feedback loop for precision farming.
How we built it
We leveraged a diverse technology stack to build this AI-powered platform. The front-end was developed using a modern framework , React to create a responsive and intuitive user interface. For the backend, we used SpringBoot . The AI for disease detection was a key component, built using PyTorch and trained on a large dataset of crop images. We integrated a third-party weather API to pull real-time forecasts. For the crop advisory and market prices features, we built a data pipeline to scrape and process agricultural data, which is then analyzed by our algorithms. The soil monitoring functionality was built with an API that can receive and display data from simple IoT sensors, simulating a real-world connected farm. The entire application is deployed on a cloud platform like Google Cloud to ensure scalability and reliability.
Challenges we ran into
The primary challenge was sourcing and curating a high-quality dataset for our AI disease detection model. We needed a diverse collection of images for various crops and diseases to ensure accuracy. Another major hurdle was integrating multiple external data sources for weather, market prices, and soil data, all while maintaining a consistent and reliable user experience. We also faced the challenge of making a complex platform accessible to users who may have limited technical skills. This required extensive user experience (UX) design and testing to simplify the interface and make the features easy to understand and use.
Accomplishments that we're proud of
We are most proud of successfully integrating multiple AI and data-driven features into a single, cohesive platform. Building a working disease detection model and a dynamic crop advisory system within the hackathon's timeframe was a significant achievement. We are also proud of our user-centric design, which makes the platform's advanced functionalities approachable for small-scale farmers. This pivot to an AI-driven model represents a more impactful and scalable solution that can truly transform farming practices.
What we learned
Through this pivot, we learned that providing a direct, data-driven service to farmers is more impactful than simply connecting them to markets. We gained valuable experience in developing and training machine learning models for real-world applications and integrating various APIs and data sources. We also reinforced the importance of building with the end-user in mind, focusing on simplicity and clarity in every feature. This project taught us that technology can be a powerful tool for social good, provided it is designed to address a core problem at its root.
What's next for Twala
Looking ahead, we plan to expand Twala to include a more robust and personalized advisory service, potentially integrating with messaging platforms like WhatsApp to provide information directly to farmers' phones. We also aim to collaborate with agricultural experts and research institutions to continuously improve our AI models and the quality of our recommendations. Our ultimate goal is to evolve Twala into a comprehensive ecosystem that includes not only advisory services but also financial inclusion features, like access to micro-loans and insurance, all powered by the data we collect and analyze.
Log in or sign up for Devpost to join the conversation.