A sneak peak at Airbnb business using useful Python data analysis tools

Since its first appearance in the market in 2008, Airbnb has been rapidly growing while providing supplements to traditional hotels.

Airbnb is actually a platform business where both demand and supply side use service as customers by exchanging resources — money for room or vice versa. This differentiates Airbnb from traditional lodging services where demand and supply side directly interact.

Source : Business Model Toolbox (https://bmtoolbox.net/stories/airbnb/)

In order to expand my understanding further, I deep dived into Airbnb data that I obtained from Kaggle. For this blog, I summarized my analysis into key questions related to the real-world context of the data, specified as below :

  1. How well Airbnb business performed in 2016?
  2. How much growth potential did Airbnb have?
  3. Which neighborhood has more expensive listings?

Two major cities in the U.S. are sampled for this study — Seattle and Boston — and the dataset includes information about Airbnb activities in 2016.

If you are ready, let’s jump right into my discoveries!

1a. How well Airbnb business performed in 2016?

Firstly I adopted two key business metrics that hospitality analysts frequently use, which are occupancy rate and average daily rate.

Here are a brief definition of occupancy rate and average daily rate :

  • Occupancy rate: the number of rooms sold / total number of rooms listed for sales for a given day
  • Average daily rate (ADR): average price of available rooms for a given day (in the dataset, the price information was not readily available for rooms that were sold out, which I therefore excluded from calculation.)

The average daily occupancy rate in Seattle is around 67% which is about 18% points higher than the one in Boston (49%). On the other hand, the median average room rate of Boston is around US$ 196, which is US$60 higher than that of Seattle.

[Note] It is a personal preference to use median for average room rate comparison in order to avoid the effect of outliers, for example, the rental price of the entire luxury condo will be extremely high and needs to be ignored.

1b. How about RevPar?

RevPAR, or revenue per available room, is one of the key performance metrics used in the hotel industry. More detailed information can be found in the following Wikipedia link.

The calculation of RevPar is quite straightforward :

RevPar = Occupancy rate x Average daily rate, where occupancy rate is number of rooms rented / total number of rooms listed for sales

So now let’s look at the RevPar chart:

RevPar is slightly higher in Boston than in Seattle, attributed mainly to the higher level of property price.

2. How much growth potential did Airbnb have?

I would like to evaluate Airbnb’s growth potential by measuring the number of new hosts. The logic is that the more attractive is the Airbnb business, the more likely new hosts join and provide listings, which then will lead to market growth especially in supply side.

I grouped total number of new listings by year that the very first hosting happened by a unique host (please note that I wrangled the data to avoid duplicated hosts). Both Seattle and Boston recorded rapid growth in new property supplies since the establishment of Airbnb in 2008.

What is surprising however is the sudden drop in the number of new listings for 2016. One of the reasons may be an error in data collection and the provided data does not provide a comprehensive look.

Another possible reason may be found from the external factors. As of the early 2016, there was a concern over Airbnb affecting the local housing market affordability and some political actions were expected. In the meantime, Airbnb announced they could start collecting taxes in Washington state which may be able to explain the more extreme decline in new supplier in Seattle.

Sources are provided below if you are interested:

3. Which neighborhood has more expensive listings?

Here, I used Python’s GeoPandas libraries to conduct geographical mapping for visualization.

The two U.S. cities share the similar characteristics, where listing prices are the highest near the central downtown and bay area. As properties are located farther away from the central area, the price decreases.

Price maps are attached below for Seattle and Boston. Just for the purpose of the completeness, the top 10 most expensive areas are listed together with the map.

Top 10 Most Expensive neighborhoods in Seattle 
Based on Airbnb listing price, 2016
NEIGHBORHOOD US$
Southeast Magnolia 231.71
Portage Bay 227.86
Westlake 194.47
West Queen Anne 187.77
Montlake 182.79
Briarcliff 176.57
Sunset Hill 176.06
Industrial District 173.33
Alki 171.62
Windermere 169.90
Top 10 Most Expensive neighborhoods in Boston 
Based on Airbnb average listing price, 2016
NEIGHBORHOOD US$
South Boston Waterfront 306.06
Bay Village 266.83
Leather District 253.60
Back Bay 240.95
Downtown 236.46
Chinatown 232.35
Beacon Hill 224.44
Fenway 220.39
West End 209.59
South End 204.35

Conclusion

In this post, we quickly explored Airbnb business by looking at real-word data for two major cities in the U.S.

Key findings are:

  1. General Airbnb performance is slightly better in Boston than in Seattle, when measured by RevPar, which was contributed by higher average rental price.
  2. Since the establishment in 2008, the growth of Airbnb was rapid in both Seattle and Boston areas. However, there was a sudden drop in the number of new listings for both cities, which may possibly be resulted from some external effects (i.e. politics or tax) supposing data collection was complete. This, however, leaves a room for validation from future studies.
  3. Airbnb listing prices are the highest near downtown and bay area for both cities. However, the price gets lower as properties are located farther from the central area.

Are you interested in doing a similar analysis?

In order to see this analysis in more detail, or want to find sources of data, please see the link to my Github available here.