AirBnB offers many tourists and travelers a relatively affordable and convenient accommodation option and home owners an extra source of income. This project aims to estimate the effects of the property characteristics, the time left until the booking starts and the season on the price per night of Hawaiian AirBnB listing by using the publicly available data from Inside AirBnB.
Our findings are not only relevant for tourists by giving them insights into how prices fluctuate across time or which neighborhood is the cheapest, and thus help them save money on their trip, but also AirBnB hosts by helping them develop a suitable pricing strategy based on the characteristics of the properties they own.
What are the effects of the property characteristics and the number of days left until the booking starts on the price per night of Hawaiian AirBnB listings and do prices differ per season?
Figure 1. Conceptual Model
Overview of the variables included in this study: | Variable (group) | Description |
---|---|---|
room_type | Since there are 4 room types: "Entire home/apt", "Private room", "Shared room", "Hotel", three dummy variables have been coded (a hotel room being the baseline) for the purpose of including them in the regression analysis. Entire_home_apt_room_type (1 if listing is an entire home or apartment, 0 if the listing is of a different type); Private_room_room_type (1 if listing is a private room, 0 if the listing is of a different type). Shared_room_room_type (1 if listing is a shared room, 0 if the listing is of a different type) | |
neighbourhood_goup | Since in Hawaii there are 4 neighborhood groups: "Hawaii", "Kauai", "Maui", "Honolulu", three dummy variables have been coded (Honolulu being the baseline) for the purpose of including them in the regression analysis. Hawaii_neighborhood (1 if the listing is in the Hawaii neighbourhood, 0 if the listing is in a different neighbourhood. Kauai_neighborhood (1 if the listing is in the Kauai neighbourhood, 0 if the listing is in a different neighbourhood. Maui_neighborhood (1 if the listing is in the Maui neighborhood, 0 if the listing is in a different neighborhood. | |
accommodates | the maximum number of guest the property can accommodate | |
bathrooms | the number of bathrooms of the property | |
bathroom_type | dummy variable to indicate whether the bathroom(s) are private or shared | |
bedrooms | the number of bedrooms of the property | |
beds | the number of beds of the property | |
review_score_rating | the review score rating of the property (on a 5-point scale) | |
instant_bookable | dummy indicating whether the property can be booked automatically or the host's approval is needed | |
time_diff | the number of days left until the booking starts | |
winter | dummy variable indicating the season (Hawaii only has two seasons) taking the value 1 for winter (November-April) and 0 for summer (May-October) |
Overview of other variables used to merge different datasets or aggregate the data: | Variable | Used for: |
---|---|---|
id | merging the listings and the calendar datasets | |
date | aggregating the dataset per season |
Data dictionary of the raw datasets can be found here.
In order to estimate the effect of several metric and non-metric variables (property characteristics, time until booking starts and season) on another metric variable (listing price of the Hawaiian Airbnb listing), we opt for a regression analysis. We are interested in which property characteristics drive the price up or down and which have the largest effects. All independent variables are outlined in the table above. The regression equation can be summarized as follows:
_Y = β0 + β1 entire_home_apt_room_type + β2 private_room_room_type + β3 shared_room_room_type + β4 hawaii_neignborhood + β5 kauai_neignborhoodbathroom_type + β6 maui_neignborhood + β7 accommodates + β8 bathrooms + β9 bathroom_type + β10 bedrooms + β11 beds + β12 review_score_rating + β13 instant_bookable + β14 timediff + β15 * winter + ε
├── README.md
├── data
├── gen
│ ├── analysis
│ ├── data-preparation
│ └── paper
└── src
| ├── analysis
| ├── data-preparation
| └── paper
└── make file
Please follow this guide to install R.
Also, make sure you install the following packages:
install.packages("tidyverse")
install.packages("utils")
install.packages("stringr")
git clone https://github.com/course-dprep/what-influences-AirBnB-prices.git
make
The regression output can be found below.
Figure 2. Regression Output
All else equal,
Travelers looking to save money seem to be best off if they book their AirBnB approximately 90-100 days in advance (~ three months), with an average price for an entire home/apartment of around $310 and an average price for a private room of around $270. Interestingly, prices seem to spike around 200 days prior to the booking start date, with the average entire home or apartment reaching $355, making it very disadvantageous for travelers to book around six months in advance. Rather, it would be better to book a bit later.
Regardless of the room type travelers are looking for, Hawaii and Honolulu seem to be the cheapest neighborhoods and Muai is the most expensive one. Interestingly, a private room in Maui is more expensive than an entire home or apartment.
Our findings are relevant for both tourists and travelers as well as AirBnB hosts.
For hosts:
For consumers: