A host is trying to enter the Airbnb business in Seattle, can you provide him/her with a housing and pricing strategy that can maximize the revenue?
How does this help Airbnb?
- Maximizing hosts' return means maximizing Airbnb's commisions
- The code can be a core of a recommendation system that suggests pricing and policy for a new-coming host
- The code can give Airbnb insight information on what category of potential host to reach out for marketing
Using the Kaggle's Airbnb Seattle dataset, I will perform EDA based on the host listing's:
- House properties: number of rooms, beds, apartment type, locations, etc.
- Host policies: number of extra guess allowed, cancellation policy, verification from rental, etc, as well as host self-verification
- Pricing strategies: price per person, cleaning fee, price surging over the year, etc.
A light gradient boosting machine is used with 5-fold cross-validation and MAE loss for revenue predictions.
The detailed analysis can be found in this notebook
A summary presentation can be found here
