Team WomemCodeToGreen - AIWH081

Inspiration

In India, many farmers face severe economic distress, with 40-50% burdened by debt and 70% earning below the national average income. This economic hardship, coupled with high rates of farmer suicides, highlights the urgent need for effective support and policy interventions to improve farmers' financial stability and well-being.

What it does

Our model provides actionable crop selection insights by analyzing soil and market data. By inputting soil nutrients, location details, and investment information, the model predicts the best crop types to grow. Advanced machine learning techniques align these recommendations with market demand, helping farmers make informed decisions to maximize yield and profitability. Additionally, the model includes a chatbot that offers real-time assistance, helping farmers learn and apply modern agricultural practices.

How we built it

We used a combination of machine learning algorithms, including Random Forest and Support Vector Machines, to analyze and predict optimal crop choices. We also integrated a LLM-based chatbot to provide real-time assistance and guidance on modern agricultural practices. Integration challenges were addressed by ensuring data compatibility and accurate feature encoding.

Challenges we ran into

Data Compatibility Issues: Ensuring correct formatting and alignment of soil and regional data. And ensuring the dataset, that is accurate for building chat bot. Feature Encoding: Converting categorical data into numerical values without introducing biases. Building CNN: Building CNN was been difficult. Communication: Communicating with team members who are present in different parts of country.

Accomplishments that we're proud of

Successfully built and implemented LLM and machine learning models for early crop prediction. Achieved effective team collaboration across different states, overcoming challenges and completing the project despite tough circumstances. Collected and utilized diverse datasets for model training.

What we learned

Building and optimizing LLMs for agricultural datasets. Implementing CNNs for data analysis. Enhancing team collaboration and problem-solving skills.

What's next for AgriHelper

Continue enhancing the LLM for more accurate and personalized advice. Expand dataset collection for improved model performance. Develop additional features and integrations based on user feedback and emerging need. Finally, to integrate all modules in flutter app.

Built With

• cnn
• lang-chain
• llm
• machine-learning
• python
• streamlit