Analysis of user preference and expectation on shared economy platform: An examination of correlation between points of interest on Airbnb. (June 2020)

• Record Type: Journal Article
• Authors: Abdar, Moloud; Yen, Neil Y.
• Is Part Of: Computers in human behavior. Volume 107(2020)
• Publication Date: 2020-06
• Subjects: Sharing economy -- Behavior analysis and prediction -- User preference -- Points of interest (PoI) -- Things-to-Do -- Airbnb; Interactive computer systems -- Periodicals; Man-machine systems -- Periodicals; 004.019
• DOI: 10.1016/j.chb.2018.09.039
• Languages: English
• ISSNs: 0747-5632

Abstract

Abstract: Numbers of sharing economy services are significantly increasing recently. Using computational techniques to obtain customer's behavior on these services are very helpful to understand user's preference, needs and expectation. Airbnb is one of such kinds of services that concentrates on providing a peer-to-peer short-term accommodation rental service. Instead of accommodation rental, Airbnb is also featured by its participatory environment that prompts users to interact with each other. This study attempts to explore the implicit meaning of Points of Interest (PoI) suggested by users (especially hosts), and finds out the correlations in between recommended spots and user preference. To this end, the research computes the distance of each point from center of city and main subway station of the city. Three metrics are designed: Popular Spots, Top-N Spots, and Geographically Average Spots for average distance. Statistics demonstrate that most of Points of Interest (PoI) those recommended by users have distance more than 3000 m from center of city and main sub-way station of the city. This means that even though the distance is very important factor, but however, some users do not attention to the distance from public points. Moreover, the results show that 'Food Scene' is the most popular recommended place by users on Airbnb. Highlights: An examination of correlation among Points of Interest (PoI) is investigated. The model is applied on two data sets related to 12 well-known European cities. The distance from center of city and main subway station of the city for each PoI is evaluated. The common types of recommended places for each city are evaluated. A cluster-based model is proposed based on distance of each point.