Forecasting climate change using deep learning in Keras

This repository starts with loading a raw dirty dataset which requires lots of data cleaning and preprocessing. After performing necessary feature engineering steps and exploring the data with visualizations, data is prepared to feed into LSTM model. After that, LSTM models are built, trained and evaluated for tasks: predicting global temperature trend and predicting US temperature trend.

To summarize, this repository performs following operations:

1. Loading dataset
2. Performing exploratory data analysis
3. Performing data cleaning
4. Performing data visualization
5. Preparing data to train the global model (global temperature trend)
6. Building and training LSTM model for predicting global temperature trend
7. Evaluating the performance of the model built in step 6
8. Preparing the data to train the US model (US temperature trend)
9. Building and training LSTM model for predicting US temperature trend
10. Evaluating the performance of the model built in step 9