Maximizing Profits for Airbnb Hosts in Hawaii
The skills the authors demonstrated here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
As businesses grow and become more popular, it is imperative that they learn to adapt their model in a way that does not negatively effect local environments. Since I am a Data Scientist, I am committed to both business and environment so that they may find ways to live together symbiotically.
Introduction & Purpose
Airbnb is a popular rental service that allows individuals to make extra cash off of unused spaces in their homes. The service is popular to tourists because it offers luxurious, private accommidations for a discounted rate compared to local hotels. Airbnb makes anywhere from 3% per booking from hosts and 6 to 12% from guests.
While tourists have enjoyed utilizing Airbnb for travel, many local communities have suffered due to the influx of people. Hawaii is one of these states where the negative effects of un-restricted vaction rentals is quite apparent. The increase in tourism around small, previously rural communities has increased cost of living, which is highly undesirable for a state where the residential population makes less than $80,000 (1) a year and the land value is above $700,000 (2). Thus, the increased property value has left properties on the market for extended amounts of time. These properties are eventually bought by wealthy non-residents and transformend into vacation rentals, which displaces locals.
Fighting back
This year, Hawaii attempted to pass legislation that would restict Airbnb rentals. Though the legislation failed, it is up for review again and Hawaii will soon join the handful of states that have passed such legislation.
With this increasing trend in legislations, it is crucial for Airbnb to adapt to these new changes to keep profits steady. The purpose of this analysis is to explore the factors: rental age, total rental reviews, and total featured amenities, and their effects on nightly rates for rentals across the state of Hawaii. Using data from the Airbnb Hawaii website that includes 275 rentals in Hawaii, simple linear regressions (SLRs) and a multivariate linear regression (MLR) are used to shed light on ways hosts in Hawaii can increase their profits in the new environment that restictive legislation will eventually create in the state.
Data Collection
Data were collected from the Airbnb Hawaii website using the Selenium web scraping package in Python. When you enter the Airbnb Hawaii website, the properties appear in list form that looks like the following.
Each property contains a url on the main page that the Selenim code had to recognize and click on to access the property-specific pages. Factors of interest obtained from each property's page were:
- Rental location (Island location of rental)
- Rate (in USD)
- Total reviews
- Featured amenities
- Rental age (in years)
Below is an example of a property's page.
Data Analyses
Analyses were conducted in R Studio and Numpy in Python. Plots were generated using Seaborn. A correlation matrix was produced to examine possible relationships between nightly rates and the other factors of interest.
Nightly rates across the state of Hawaii showed a positive correlation with featured amenities (r = 0.36) and a negative correlation with total rental reviews (r = -0.46), and rental age (r = -0.09). These relationships are demonstrated visually in the plots below.
SLRs were performed to analyze the relationship between rate and each individual factor. Significant relationships were observed between rate and featured amenities (p < 0.0001) and total rental reviews (p < 0.0001). The relationship between rate and rental age was not significant.
MLR was then performed to observe relationships controlling for all other factors. In the model, it shows total rental reviews, rental age, and number of featured amenities. The MLR showed a significant negative linear relationship between total rental reviews and rate ( p < 0.0001). Significant positive linear relationships were observed between rate and featured amenities ( p < 0.001). All other relationships were not significant.
Plots
Conclusions and Recommendations
Nightly rental rates shared significant linear relationships between number of featured amenities and total rental reviews. Based on the analysis, for a regular host or Superhost to increase their profit, they should increase the number of featured amenities they offer and decrease their rate. If hosts decreases their rate, they will receive more rental reviews, based on the relationship between the two. The assumption is that a rental with low rates and a high number of amenities will be perceived by guests as a great value. This perception could increase total number of reviews for a rental and increase its star-rating, thus increasing a rental's popularity.
In the future, I will perform more in-depth analyses on specific amenities that are associated with the highest reviews and highest rates. I will also explore the relationships observed in this analysis by island.
Sources and Links
(1) Department of Numbers (2017). https://www.deptofnumbers.com/income/hawaii/
(2) KITV4 News (2017). https://www.kitv.com/story/35822780/oahu-single-family-home-prices-vault-to-all-time-high
To examine the code for this project, click the link to my project Github: https://github.com/laurajel/Airbnb-Web-Scrape
About Author
