How to reproduce and extend this work

To reproduce and extent this work, go through the following steps.

1. Get a GitHub account (it’s free).

2. Fork this repository. That is - make your own copy of it in your GitHub account.

3. Clone your fork to your local machine.

4. (You don’t have to do the steps above if you just want to download the code and data to run it locally. You can download the whole thing as a zip file. But I do strongly recommend them - it makes it much easier to keep track of what you’ve done, and to share your work with others.)

5. Set up the necessary software environment.

6. Run the scripts to reproduce the work:

   1. Download the data

   2. Normalise the data

   3. Train the model

   4. Evaluate the model

   5. Assimilate temperature and pressure to calculate precipitation

7. Now you are ready to extend the work. You can try:

   1. Retraining the model on different data.

   2. Changing the model hyperparameters - maybe training rate or beta.

   3. Changing the model: More layers? Different activation functions? Different loss function? More features?

   4. Whatever - be creative. And share your results!