EcoGuardian: AI-Driven Health Advisor

Inspiration

Air pollution isn't just a number; it's a silent health crisis. Most existing platforms show complex chemical concentrations that common citizens don't understand. We were inspired to build EcoGuardian to bridge this gap—translating raw sensor data into immediate, life-saving medical advice.

What it does

EcoGuardian is a dual-purpose AI system designed for environmental monitoring and public health safety. Predictive Analytics: It processes raw concentrations of harmful pollutants including SO_2 (Sulfur Dioxide), NO_2 (Nitrogen Dioxide), and Particulate Matter (RSPM/SPM) to calculate a real-time Air Quality Index (AQI). Medical Translation: Unlike traditional trackers that only show numbers, our system uses a custom-built intelligence engine to translate data into Actionable Health Advice. Risk Categorization: It automatically classifies the environment into safety zones (Good, Satisfactory, Moderate, or Poor) and provides specific precautions for high-risk groups, such as heart and lung patients. Deployment Ready: The system is designed to be lightweight, allowing it to be integrated into smart city dashboards or low-cost IoT sensor networks for hyper-local monitoring.

How we built it

We developed a predictive intelligence layer using: Machine Learning: Implemented a Random Forest Regressor to handle the non-linear nature of environmental pollutants. Data Processing: Used Pandas and NumPy to clean raw Indian AQI datasets, handling missing values and encoding issues (Latin-1). Heuristic Engine: Created a logic-based system to map predicted AQI to specific healthcare recommendations for vulnerable groups.

Challenges we ran into

One of the biggest hurdles was dealing with inconsistent data types and missing values in raw sensor logs. We overcame this by implementing a robust preprocessing pipeline that standardizes pollutant inputs like SO_2, NO_2, and Particulate Matter (RSPM/SPM).

Accomplishments that we're proud of

Successfully training a model that provides reliable predictions from raw industrial data. Creating a professional output interface that makes data "human-readable." Developing a solution that is lightweight enough to be deployed on edge devices.

What we learned

We deepened our understanding of Ensemble Learning and the critical importance of data encoding in real-world datasets. We also learned how to translate technical metrics into social impact.

What's next for EcoGuardian: AI-Driven Health Advisor

We aim to integrate this model with live IoT sensors for hyper-local monitoring and develop a mobile app that sends automated health alerts to patients with respiratory conditions based on their specific GPS location.

Built With

• google-colab
• numpy
• pandas
• python
• scikit-learn