MLINPrediction/pollution_prediction

Repository containing the application for the Machine Learning in Industry Practical

The basic structure of the repository is as follows, from the root folder:

• —data/: Folder with the prediction datasets.

• —preprocessing/: Folder with the code necessary for API fetching, working the data and feature engineering.

• —README.md: File with instructions.

• —requirements.txt: File with the dependencies.

CODE INSTRUCTIONS

Running application

Running "streamlit run app.py" builds an interactive application for predicting O3 and NO2 levels. The pretrained models are loaded and predictions are made every 6 hours with the new, updated data. The predictions are saved under "data".