ML bootcamp: from idea to MVP

This is a project for ML bootcamp by HSE: exploratory data analysis, modeling and building a service.

Selected topic: flight satisfaction. Dataset: here

Files:

Flight_satisfaction.ipynb: Colab notebook with exploratory data analysis, modelling and interpretations
app.py: Streamlit app file
model.py: script that transforms data, trains and runs the model
requirements.txt: package requirements files
/data folder has:
- clients.csv: copy of dataset
- importances.csv: features sorted by their importance
- model_weights.mw: pretrained model
- and png visualizations for EDA

Streamlit service is available at rateyourflight.streamlit.app via Streamlit Cloud

To run locally, clone the repo and do the following:

$ python -m venv venv
$ source venv/bin/activate
$ pip install -r requirements.txt
$ streamlit run app.py

The app will be available at http://localhost:8501