ML-driven dynamic pricing models for revenue management of short-term accommodations

With the fast development of machine learning (ML) technologies, it is important to understand how they can help businesses optimize their revenue. In this thesis, the focus is on short-term accommodation, such as Airbnb provides. This research paper aims to analyze a dataset of 3551 Airbnb listings in Amsterdam to find what features of a listing influence its price the most and what external factors such as seasonality and events in the city have an effect on the price. Additionally, this paper describes what benefits ML techniques could bring to the pricing dynamics and market equilibrium of the short-term rental industry. The key findings revealed that numerous property features, including the accommodation capacity, location, and number of reviews, significantly impacted the pricing of a listing. Additionally, host attributes, such as their years of experience on the platform and superhost status, also played a crucial role in determining the price. In terms of external influences, seasonality was found to be much more influential on the average daily price of listings compared to the local events.