Air pollution monitoring and prediction using IOT and ML

Inspiration

With the ever increasing air pollution levels which affects millions of people around the world, we decided to base our project on monitoring and predicting Air pollution to provide critical insights to the levels of Air pollution and take active measures to reduce it.

What it does

A system which monitors the air quality using MiCS6814 sensor, MQ135 sensor, MQ131 sensor and PM2.5 sensor and forecasts the Air Quality Index for next five hours using Linear Regression, Support Vector Regression, SARIMAX model, Gradient Boosted Decision Tree ensemble model and Stacked ensemble model is proposed. The proposed Machine Learning models are compared using RMSE as a metric and the model with lower RMSE value is chosen. This project can be used in major cities to monitor the air quality remotely and can in turn help to reduce the air pollution level.

How we built it

The IOT part was built using an MiCS6814 sensor, MQ135 sensor, MQ131 sensor, PM2.5 sensor, Arduino and NodeMCU. The data from the IOT sensors are stored in the cloud in form of a Google Sheet and the Machine Learning models are applied to this dataset to forecast the AQI of next five hours and the Root Mean Squared Error is used as a metric to find the model which is most accurate.

Challenges we ran into

One of the major challenges we faced was there weren't as many references for forecasting in Machine Learning as there were for prediction in Machine Learning. Also, we faced difficulties procuring sensors for the project in the middle of a global pandemic. Another challenge we faced was for the Stacked ensemble model there weren't any references to forecast the AQI for multiple hours but we found a way to make it work.