SmogSenseAI: Our 48-hour Hackathon Project
SmogSenseAI: Our 48-hour Hackathon Project
I am thrilled to share the successful completion of a 48-hour hackathon as part of the HEC Generative AI Training (Cohort 3) organized by Pak Angels and iCodeGuru. Our team developed SmogSenseAI, an AI-powered platform designed to tackle the complexities of environmental data automation and smog analysis.
🔍 The Challenge
Environmental research papers are often unstructured and dense, making it incredibly time-consuming for researchers to extract critical pollutant data like $PM2.5$, $NO_x$, and $SO_2$.
💡 Our Solution: SmogSenseAI
We built an end-to-end system where users can upload complex research PDFs. The AI then:
Generates a concise executive summary.
Extracts key pollutant metrics accurately.
Creates structured, actionable analysis reports for better decision-making.
🛠️ Tech Stack & Contribution
Working on the backend logic, I focused on integrating powerful APIs to ensure the system delivered precise insights. Our stack included:
Language: Python
Framework: Streamlit
LLM Integration: Groq API
Data Extraction: pdfplumber
Deployment: Hugging Face
It was an incredible experience collaborating with my team to bring this idea to life. Looking forward to applying these Generative AI skills to solve more real-world problems! 💻✨


