The problem Udyan Saathi solves

Our team, Tech Tonics, comprises four BTech CSE undergraduates from Bennett University: Vansh Mehta, Mannan Tyagi, Abhay Chauhan, and Yashvir Singh. We are developing UDYAAN SAATHI, a web-based environment monitoring system that provides real-time AQI (Air Quality Index), WQI (Water Quality Index), and weather monitoring and forecasting across India.

Our solution leverages government APIs to fetch raw data, which is processed through a multi-layered pipeline using Python Pandas:

Bronze: Raw data ingestion
Silver: Data cleansing and formatting
Gold: Data transformation and AQI calculation
Platinum: Consolidation of data for machine learning models
The process is automated via a Windows service that runs daily. Our platform visualizes and compares data using heat maps, bar graphs, and line graphs. We use historical data for comprehensive insights and employ the GRAP (Graded Response Action Plan System) to alert users and authorities about harmful levels of toxins, offering actionable plans. Our ML model forecasts data with an R-squared score of 95-98% and a Mean Absolute Error of 3-5.

To handle big data and optimize our application, we utilize Azure Cloud Services:

Azure Data Factory: Orchestrates data workflows and schedules pipelines
Azure Data Lake Storage Gen2: Stores raw and transformed data
Azure Databricks: Provides computational power for running PySpark notebooks and ML algorithms
Azure MySQL: Saves the final formatted data
Azure Web App Service: Hosts the Django REST API Service
Azure Key Vault: Stores private keys and connection strings for enhanced security
Azure Static Web App Service: Hosts the user interface layer

Challenges we ran into

We Ran into Various challenges
Creation of Complex Pipeline on our local system to process large dataset and automate it using Windows Service
Storing Data in MySql Tables and Fetching data form them and integrating it in frontend
Shifting the project on Azure we faced multiple challenges during deployment