PRESENTATION
Covid-19 fourth wave it's close in Italy, but how we can know when the situation starts to be critical?
Since past events we know that the new cases growth is almost an exponential, when a wave starts, so we want to predict the future trend, with a good model, knowing what happened in previous days.
We also know that new cases, intensive care unit admissions and deaths are strictly correlated. When one increases, also the others are increasing in the same order of magnitde.
Building a good model is essential in order to avoid taking bad decisions (for example lockdown when not needed, or no restrictions when needed).
Obviously the prediction is really complicated, because a lot of variables have to be take into account.
An example: the third vaccination may produce a decrease effect in future trend (especially deaths).
As you may understood by the previous lines, the aim of this project is finding a good model to predict future trend, using statistical learning theory.
The regression functions that have been used are the following one:
All data are public and provided by the Dipartimento della Protezione Civile (DPC).
You can also find them at this github link.
Before downloading the source code, you should enable the dark theme on your jupyter notebook, in order to see graphs and other items in a proper way.
So you have to run the following lines of code in your terminal/command prompt:
pip install jupyterthemes
jt -t chesterish
Finally, you have to run the following code snippet, in order to get all the libraries that have been used in the project:
pip install DateTime
pip install ipywidgets
pip install python-math
pip install matplotlib
pip install numpy
pip install pandas
pip install patsy
pip install -U scikit-learn
Are you interested about this argument?
- Read this beautiful book: The Elements of Statistical Learning.
- You can also see another project: Gas consumption prediction