• 

• 

• 

• 

🌱 FarmSage: Crop Recommendations for Sustainable Farming 💡 Inspiration Agriculture is a lifeline for millions, but choosing the right crop for a specific location can be difficult due to changing climate, soil degradation, and inconsistent rainfall. We were inspired to create a solution that empowers farmers with intelligent, data-driven decisions, allowing them to choose the best crop based on measurable soil and weather parameters.

By leveraging machine learning, we aim to contribute to more sustainable farming practices, better yields, and improved resource usage.

🤖 What it does FarmSage is a crop recommendation system that:

Takes soil nutrient values ((N), (P), (K)), pH, rainfall, temperature, and humidity as inputs

Uses a trained machine learning model to predict the most suitable crop for the given conditions

Offers a user-friendly web interface built with Streamlit for accessibility

🛠️ How we built it Dataset We used a public dataset containing features like:

Nitrogen ((N))

Phosphorus ((P))

Potassium ((K))

Temperature

Humidity

pH

Rainfall Each row was labeled with the ideal crop.

Preprocessing

Normalized and cleaned the dataset

Split into training and testing sets using train_test_split

Model Training Tested several classification algorithms:

Logistic Regression

Decision Tree

Support Vector Machine (SVM)

Random Forest Random Forest gave the best performance:

python Copy Edit from sklearn.ensemble import RandomForestClassifier model = RandomForestClassifier() model.fit(X_train, y_train) Evaluation We used metrics like accuracy and confusion matrix:

Accuracy

𝑇 𝑃 + 𝑇 𝑁 𝑇 𝑃 + 𝑇 𝑁 + 𝐹 𝑃 + 𝐹 𝑁 Accuracy= TP+TN+FP+FN TP+TN ​

Deployment Built an interactive Streamlit web app where users can input data and receive crop recommendations instantly.

🚧 Challenges we ran into Data limitations: Agricultural datasets often lack diversity or real-time updates.

Model overfitting: Balancing performance on both training and testing data was tricky.

UI/UX for Streamlit: Designing a simple and intuitive form took time and iteration.

Feature scaling: Ensuring consistent scale across variables to prevent bias in prediction.

🏆 Accomplishments that we're proud of Built a fully functional crop recommendation system

Achieved high model accuracy (above 95%) with Random Forest

Successfully deployed the system with an intuitive interface

Translated complex machine learning predictions into real-world utility

📚 What we learned How to work with real-world agricultural data

Importance of data preprocessing in model performance

Strengths of various classification algorithms

Deployment skills using Streamlit

How technology can genuinely impact sustainable agriculture

🚀 What's next for FarmSage: Crop Recommendations for Sustainable Farming 🔄 Real-time weather integration using APIs

📱 Develop a mobile app version for broader reach

🧪 Add fertilizer and pesticide recommendations

🗺️ Use geolocation data to personalize predictions

🧠 Explore deep learning for multi-crop prediction and precision farming

Built With

• decision-tree
• etc.)-data-preprocessing-model-evaluation-streamlit-?-for-deploying-the-machine-learning-model-in-a-user-friendly-web-interface-jupyter-notebook-?-for-prototyping
• git
• jupyter
• matplotlib
• notebook
• numpy
• pandas
• python
• scikit-learn
• seaborn
• streamlit
• svm
• visualizing-data