India Air Quality Analysis
This project performs exploratory data analysis on India's Air Quality dataset to understand patterns and trends. It is also using random forests, LSTM, and CNN architectures to forecast future quality levels. This project has the potential to make a significant contribution to the understanding and management of air pollution in India.
Description
This project is doing exploratory data analysis on India’s Air Quality dataset to understand patterns and trends. It is also using ensemble random forests, LSTM, and CNN architectures to forecast future air quality levels. This project has the potential to make a significant contribution to the understanding and management of air pollution in India.
Exploratory Data Analysis
Firstly I grouped the data into various frequencies (day, month, year) to identify possible trends:
Next I plotted the similarities between features so that I get a better explanation on the relationships between the variables:
Finally, through a correlation matrix, I can easily visualize the correlation degree between the variables.
Time Series Forecasting
Ensemble Models
- Random Forest
- Gradient Boosting
- AdaBoost
- Histogram Gradient Boosting
- XGBoost
Deep Learning Models
- Long short-term memory (LSTM)
- Stacked LSTM
- Bidirectional LSTM
- Bidirectional Stacked LSTM
- 1D Convolutional Neural Network (CNN)
- 1D Convolutional LSTM
