A weather forecast station project

Hello i have worked on 2 datasets about the weather in France, one about weather data on the sea (Gascogne gulf) from 1996 to now, and an another one about the weather in the north west of France in 2016, 2017, 2018 (MeteoNet dataset).
My project is to make a little weather forecast station device with a raspberry pi and working with machine learning or deep learning.
For the moment i have only used models like randomforest because its seems to be the more effective models for that kind of problem. I have not tried other models yet because my computer is not very powerful so it takes time to try parameters with gridsearchcv.

Below i write the adresse of the open source dataset findable on the web (MeteoNet):


And now the package i have written on github, useful if you want to convert this dataset in data directly working with sikilearn:

I am seeking about people who are interested to improve this project or who are working on an another weather dataset. Hope my english is not too bad ! See you !