GrowRight is a Flask-based web application that leverages an Artificial Neural Network (ANN) to recommend the most suitable crop to grow based on environmental and soil parameters. This project empowers farmers, researchers, and agriculture enthusiasts with data-driven crop suggestions for optimal yield.
In the world of smart agriculture, selecting the right crop is crucial for optimizing land productivity. GrowRight uses machine learning to analyze environmental factors such as nitrogen, phosphorus, potassium, temperature, humidity, pH, and rainfall to predict the best crop to cultivate. The model is trained on the Smart Agricultural Production Optimizing Engine dataset.
- ✅ Crop Recommendation: Predicts the ideal crop based on user-input environmental data.
- 🧠 Trained ANN Model: Uses a TensorFlow-based artificial neural network with high accuracy.
- 🌐 Web Interface: Simple and intuitive frontend built with HTML and CSS.
- 📊 Data-Driven Insights: Powered by real-world agricultural data.
- 🔁 Real-Time Prediction: Instantly processes inputs and returns the crop recommendation.
| Layer | Technologies |
|---|---|
| Frontend | HTML5, CSS3 |
| Backend | Flask (Python) |
| ML Model | TensorFlow (ANN), NumPy, Pandas, Scikit-learn |
| Dataset | Kaggle - Smart Agricultural Production |
| Deployment | Railway |
- Comprehensive Exploratory Data Analysis (EDA) has been performed using Python notebooks.
- The EDA includes data cleaning, visualization, and feature correlation analysis to better understand the dataset before model training.
- Insights from EDA helped in selecting relevant features and preprocessing steps for the ANN model.
- Accuracy: 99% (0.99) on the test dataset, demonstrating excellent prediction capability.
- Below is the confusion matrix for the trained ANN model, illustrating the classification performance across different crop classes.
-
🌍 Location-Based Input
Integrate geolocation APIs (like OpenWeatherMap or MapMyIndia) to autofill temperature, humidity, and rainfall values based on the user's current location. -
📱 Mobile Optimization
Improve the frontend responsiveness using CSS media queries or frameworks like Bootstrap or Tailwind CSS to ensure seamless mobile usability. -
⚙️ Model Enhancement
- Apply hyperparameter tuning (GridSearchCV, RandomizedSearchCV).
- Experiment with deeper or alternative neural architectures like CNNs/RNNs or ensemble models (Random Forest, XGBoost).
-
🧪 Soil Image-Based Prediction
Use computer vision techniques to analyze soil images and supplement or replace manual input with real-time insights. -
🧠 Explainable AI (XAI)
Integrate explainability tools like SHAP or LIME to explain model predictions to end users. -
🌐 Multilingual Support
Translate the interface into regional languages for accessibility by farmers across various linguistic backgrounds. -
📦 Docker Support
Package the app using Docker for easier deployment across different environments. -
☁️ Cloud Deployment
Deploy the model and backend on scalable cloud platforms (AWS EC2, Railway, Render, Azure, or Heroku). -
📊 Historical Data Visualization
Let users log and view their input history and predictions via charts or downloadable reports. -
🔐 User Authentication
Add login/signup functionality so users can save profiles, get personalized insights, and track usage. -
🛰️ IoT Sensor Integration
Accept real-time data feeds from physical sensors (IoT devices) for temperature, pH, or moisture levels. -
💬 Chatbot or Voice Assistant
Integrate a smart assistant to answer agricultural queries and guide users through the app.