Exploring Airbnb’s Tokyo Data Set
Have a property in Tokyo to list in Airbnb? Perhaps data can help with the pricing
For new hosts listing on Airbnb, pricing might be a grey area. What price should they put on the listing? Is it too expensive? Is it too cheap?
This project aims to give a starter fee structure for new hosts. From the 2019 data analysis, it seems that:
- A 1-night minimum stay works great; and
- Hosts can start their price at the median of the listings
The interactive visualization can be found here.
This is a project done for the Data Analytics course with General Assembly.
For the Airbnb data set, we are given the following problem statement:
You’re an analyst working for Airbnb, and you’re doing research in order to recommend a fee structure for hosts to maximize their yearly revenue.
Fees in Airbnb website include price per night, minimum night requirements, cleaning fees, security deposits, and fees for extra guests. Considering that the data sets do not have any information on cleaning fees, security deposits, and fees for extra guests, this project will focus on the price per night and the minimum night requirements.
Data and Analysis
The data sets are available at Inside Airbnb. The files used are listings.csv.gz and reviews.csv.gz. Data handling is done on Excel and visualization is done on Tableau.
At the time the listings file was accessed on 21 Mar 2021, there were 11,308 listings. To check whether there is a demand for a particular listing, a non-automated review count is used to approximate booking. Considering the most recent numbers below, only listings that have at least 1 review in 2019 will be used.
On 15th Jun 2018, Japan’s minpaku law came into effect and tightened Japan’s regulation on short-term rentals, listings were cut down.
In 2019, Tokyo Stadium hosted some of the matches for Rugby World Cup 2019.
In 2020, tourism was badly hit due to covid-19.
Visualizing the screened data on the map, most listings are located in the 23 Special Wards. Hence, the project focuses on the listings in these areas.
The majority of these listings have a 1-night minimum stay requirement.
A quick check on the price distribution across all room types shows that the price distribution is skewed. It seems that there is a need to break the data further into the different property types.
This conclusion is based on an assumption that the prospective host has a property 1) located in a typically strategic location, 2) with a typical number of baths and beds, and 3) with typical amenities available for that type of property.
Ideal minimum nights: 1 night
Starting prices: hosts can consider starting with the median price in the box plots below.
About the author
Alden is a finance professional who is currently exploring the world of data.
You can connect with him on LinkedIn.