Reservation Cancellation Forecasting: a Senior Machine Learning Engineer assignment

In this repository, you'll find a project I handled during a Machine Learning assignment for a leader in the short-term rental industry.

The objective of this assignment was clear:

• Build a Machine Learning model to predict whether a reservation is likely to be cancelled,
• Develop an MLOps architecture designed for a production environment.

To accomplish this task, Airflow and Mlflow are used to build Machine Learning Pipelines, fully customizable and ready for a production environment.

This Github repo comes with a complete article published in the Towards Data Science journal on Medium:

Article link

Build Machine Learning Pipelines with Airflow and Mlflow: Reservation Cancellation Forecasting.

Code organization

.
├── README.md
├── airflow
│   ├── dags
│   │   ├── inference_pipeline.py
│   │   └── training_pipeline.py
├── artifacts
├── data
│   ├── features_store
│   ├── preprocessed
│   ├── hotel_bookings.parquet
│   └── sample_for_inference.parquet
├── mlflow
├── notebooks
│   ├── 0_exploratory_data_analysis.ipynb
│   └── 1_preprocessed_data_check.ipynb
├── requirements.txt
└── steps
    ├── condition_step.py
    ├── config.py
    ├── feature_engineering_step.py
    ├── inference_step.py
    ├── preprocess_step.py
    ├── train_step.py
    └── utils
        ├── _artifact.py
        └── data_classes.py

The repository is structured as follows:

• Data Exploratory Analysis (EDA) is performed on notebooks,
• Each stage of the Machine Learning process (Preprocessing, Training, Inference, etc...) is defined as a module designed to be implemented into a pipeline. They are all located in the steps/ folder.
• Airflow and Mlflow are deployed locally within this repository.
• In the data folder is located the original dataset that was provided for this assignement, in addition of a sample for batch prediction. data/features_store and data/preprocessed are directories to store the data once processed by some stages of the pipelines, such as preprocessing or features_engineering steps.
• The same idea for artifacts that contains encoders generated during the features_engineering step.

(back to top)

Getting started

The code runs with Airflow and Mlflow.

To launch these applications, open a terminal for each and type their respective command lines after having installed them. The complete procedure can be found in the article.

# Terminal 1
mlflow server --backend-store-uri mlflow/ --artifacts-destination mlflow/ --port 8000

# Terminal 2
airflow standalone

(back to top)

Contributing

Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement". Don't forget to give the project a star! Thanks again!

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

(back to top)