🌍 Inspiration

Large organizations and corporate companies are striving to increase sustainability globally. Gen AI has numerous use cases in sustainability development, one of which is Sustainability Reporting and Boosting Collaboration within companies. To enhance communication and reporting for sustainability, we developed Sustainability Analytics.

---

💡 What it Does

Sustainability Analytics provides an intelligent chatbot interface that allows users to ask real-time questions about a company’s sustainability data.

Key Features:

• Leverages ESG data and advanced AI models (like LLaMA 3.1) to retrieve accurate and relevant information.
• Users can inquire about:
  • Carbon emissions
  • Energy usage
  • Water consumption
  • And more!
• The chatbot generates:
  • Year-on-year comparisons
  • Visual insights in the form of bar, line, and pie charts.

Example:

User Query: "What are the total carbon emissions in 2024 for the respective company?"
Chatbot Response: "The total carbon emissions for 2024 are X metric tons."

---

🛠️ How We Built It

Tech Stack:

• Advanced RAG methodology using the open-source LLaMA 3.1 model.
• Langchain framework for querying data.
• Postgres database to store ESG metrics.
• Backend: Python framework FastAPI.
• Frontend: Built using React.js.

Data is queried through Langchain tools, which the LLM processes to generate natural language responses, along with chart visualizations.

---

🚧 Challenges We Ran Into

We encountered several challenges during development:

1. Prompt Engineering:

   • Ensuring the LLaMA 3.1 model accurately handles sustainability-related queries.

2. Data Sourcing & Integration:

   • Structuring ESG metrics data for easy querying through the Langchain framework.

3. Performance Optimization:

   • Enhancing the RAG methodology for large datasets and managing chart generation (bar, line, pie) for data comparisons.

4. Accuracy Across Queries:

   • Ensuring accuracy in dynamic year-on-year comparisons across diverse query types.

---

🏆 Accomplishments We're Proud Of

• Successful Integration: We integrated LLaMA 3.1 with Langchain to build an interactive sustainability analytics platform.
• Real-Time Responses: Our chatbot delivers accurate, context-specific answers about a company’s ESG metrics in real-time.
• Dynamic Visualizations: Users can generate visual insights (bar, line, pie charts) based on queries.
• Efficient Querying: Built a highly efficient querying mechanism for our Postgres database, ensuring speed without compromising accuracy.

---

🎓 What We Learned

• Mastering advanced RAG methodology helped us streamline the generation of precise responses.
• We learned to handle ESG data more effectively while building scalable backend systems that support data-heavy operations.
• Integrating databases with generative models taught us the importance of data integrity and query optimization.

---

🚀 What’s Next for Sustainability Analytics

We aim to expand Sustainability Analytics with:

1. Advanced Data Analytics Features:

   • Predictive analysis for forecasting future ESG metrics.
   • Identifying areas for sustainability improvement.

2. Global Standards Integration:

   • Adding more sustainability frameworks to align with global standards.
   • Support for multilingual capabilities for a broader client base.

3. Collaboration Tools:

   • Introducing tools that allow company stakeholders to collaboratively input data and generate comprehensive reports.

4. API Integration:

   • Expanding API integrations with existing sustainability platforms.

5. User Interface Improvements:

   • Enhancing the UI for a more intuitive user experience.

---