Aerogram (IIT Delhi)

Aerogram is an IIT-Delhi based startup under TBIU that is devising a network to predict real-time air quality in a local mapped area. Utilised the Time Series Forecasting to predict and forecast PM 2.5 values and other Air pollution metrics using Machine Learning models and regression analysis. Developed Python scripts and ETL Pipelines for publishing and subscribing to telemetry data to integrate with MQTT Architecture to send and receive Air Pollution metrics on Google Pub/Sub and Google IoT core for further analytics. Analyzed the real-time sensor data generated by the Ezio-Stat devices for fetching the PM 2.5 values and analyzing the seasonality, trend, and the noise component to further understand the various factors that contribute to the fluctuations of PM 2.5 in correlation to temperature, pressure, and humidity. Prototyped a Web dashboard by using HTML, CSS, JS, Flask, and Cloud Firestore for the end-user to forecast PM 2.5 value a few hours ahead so as to provide insights. Worked on MQTT protocols for migrating telemetry feed from low bandwidth IoT-Devices to Google IoT Core for further data collection. Developed Google Cloud Functions using Python as a programming language to migrate the telemetry feed received from sensors to Google Cloud Pub/Sub, Google Cloud SQL, and Google Cloud Firestore. Determined the weekly and monthly comparison of the Air Quality Metrics such as PM 2.5 during, before and the COVID-19 lockdown Utilized Tableau to plot the monthly and weekly comparisons of BAM and the IIT Sensors to crosscheck values with respect to Central, West, North, and East Delhi. Determined the Correlation & R^2 between the Aerogram Senors and BAM. Utilized the Scatter Plots for determining the visualization representing the Correlation. Performed feature engineering and data cleansing from raw data for performing analysis on the metrics provided by the Aerogram Sensors.