Airbnb Recommendations

The code here is for training, validating, and deploying an Airbnb Country Recommendation System, using an XGB Classifier and SkLearn packages.

What's the Airbnb Recommendations?

Our Airbnb Recommendations is a machine learning library for predicting the new countries that a new user will book an airbnb in. There are a total of 10+ countries of interest, that need to be recommended given the Airbnb dataset.

Set up

System Requirements

1. Python 3.6 or higher
2. Python libraries: pandas, scikit-learn, matplotlib, numpy, tqdm, xgb

Conda Setup Instructions

If you are using conda, you can get started by cloning this repository and using the environment.yaml file as in the following:

conda env create -n airbnb_env -f env.yml

and then activate the environment using the following,

conda activate airbnb_env

Docker Setup Instructions

If you are using docker, there is a Dockerfile included. After cloning this repository, you can build the image with the following,

docker build -t airbnb_recommender .

and then run it locally using,

docker run -it -p 8501:8501 airbnb_recommender

Download required files

Download the dataset and store it within the project dir as airbnb-recruiting-new-user-bookings. Next create the dataset. Here we will create our engineered features and process the dataset.

cd data
python make_dataset.py

Doing this, you'll output a processed baseline csv file airbnb-recruiting-new-user-bookings/train_users_2-processed.csv

Train and evaluate model

cd <PROJECT_DIR>
python train_model.py

Running streamlit application

streamlit run src/navigation_app.py

Directory Structure

The directory structure of the Airbnb Recommendations project looks like this:

├── README.md                               <- The top-level README for developers using this project.
├── data                                    <- Scripts to download or generate data
│   ├── merge_baseline_sessions.py
│   ├── d_utils.py
│   └── make_dataset.py
├── sessions-data                           <- Scripts to preprocess and feature engineer sessions data
│   ├── create-sessions-casting+ratio-csv.py
│   ├── merge-sessions.py
│   ├── pearson-features.py
│   └── generate_session_distinct_counts_and_time_features.py
├── models                                  <- Scripts to select and evaluate model
│   ├── select_model.py
│   └── eval_model.py
├── airbnb-recruiting-new-user-bookings     <- Expected data file
│
└── train_model.py                          <- Main script