AirbnBoost is an online platform powered by machine learning that enables users to make faster and better informed Airbnb decisions.
- See how each listing is priced compared to similar listings based on AirbnBoost's price prediction algorithm. Find great deals based on the deviation between the pricing algorithm prediction and the actual listing price.
- Select listings that match your preferences without the need to read through full descriptions.
- Understand the urban environment surrounding a listing, such as proximity to subway or neighborhood noise levels.
Here is a step-by-step navigation at AirbnBoost UI.
The implementation has three main components:
01 - data_preprocessing.ipynb, a notebook that geolocates and spatially joins Airbnb listing data with urban data sets and processes them before they are used as inputs in the modeling part.02 - data_modeling.ipynb, a notebook that trains the two models behind AirbnBoost. An LDA model extracts the topics from the listing descriptions. XGBoost is the algorithm behind the pricing model.app.py, is a Flask web app deployed on Heroku cloud.
Copyright © 2019 Sokratis Papadopoulos. All rights reserved.