Inside Airbnb: Analyzing Austin, TX

We all know about Airbnb, a ‘community marketplace’ where hosts can list accommodations and travelers can find a place to stay. I read an article the other day about an Airbnb host who made $41,879/year. In the article, he talks about how based on his experience of two years as an Airbnb superhost, he was able to attract people to his listing.  I thought this would be a perfect opportunity for a data analysis and visualization.

The questions I decided to answer: 

  • What factors affect the availability of home rentals?
  • What can we learn about the opportunity for rentals if we model listing data?
  • Which variables can provide the greatest value to hosts in the Austin area?

Exploring the dataset earlier showed that though the extremely high-value rentals may sometimes be actual mansions or whatnot for rent, for the most part, they’re scalpers setting unrealistic valuations. These will throw off our model if we include them, by a lot. Exploring Price showed that a price ceiling of around $1000 per day was reasonable, so I just used that. Here is an example:

Austin Tarry Town F1 Weekend Estate $10000/night (link)

 

 

Exploring Prices

I started by exploring the data, the distribution of prices of Airbnb listings in Austin, Texas in general.
1
Avg.Price vs Zipcode: As one can imagine Rentals downtown (78712) and affluent neighborhoods 787846 will drive a greater demand + price of rentals.
2
Naturally, the type of residence has a strong effect on price. Note that the vast majority of observations fall into just five categories: House, Apartment, Condominium, and Townhouse and villa. 
The Boat effect is interesting, people are paying a good price to be on the water. Also, there is an Island for rent in Austin, really?!?!?
price_vs_bed_type
More understandable effects, whether your bed is real or not.

Let’s take a look at the market in terms of bedrooms and bathroom provided. Below is a heat map showing the number of BnBs of various Bathroom/Bedroom configurations.

Bed_vs_Bath
Not surprisingly housing availability is clustered around the 1 bed 1 bath, with dorms and such having less and large rentals having more. The vast majority of entities fall into a 0x0-to-2×2 box.

Modeling Prices

I tried a simple linear regression classifier to AirBnB prices for rentals in the Austin market, keeping in mind predictor variables that were important while exploring prices.

Observance

  • The number of listings by one host have an impact on price listing.
  • According to my model, every additional bedroom will cost $49, also it was very interesting to see that every additional bathroom will cost additional $88.
  • Renters going through Airbnb, by and large, are people looking for a relatively cheap short-term place to stay
  • Each additional bed is just $5 dollars, and each additional accommodation is $16.
  • Surprisingly, the number of reviews decreases price by 69 cents.
  • Host review ratings increase price by $2

Neighborhood Effect

6
We see that neighborhood does indeed, unsurprisingly, have a strong effect on the price of a listing. Far from downtown Austin, such as zip code 78728 (Pflugerville) almost $85 or less than average to rent in, which as a lifelong commuter honestly sounds like kind of a good deal to me (hey, it’s still Austin). Whereas right next to UT campus (78712) is the premium location.

Amenities Effect

Amenities are an interesting case.

  • On a face level, more amenities are always better: Hence on an individual level, there’s an incentive for a host to list irons, hot tubs, kitchens, the works.
  • So the effect of amenities in our model should be interpreted not as a fungible good but as a social signal. Every additional amenity available at a location will theoretically be a big plus for people who actually want it and a not-negative for those who don’t (you can choose, after all, whether or not they make use of it).
  • A price penalty for including a kitchen, hangers, or Internet access makes sense, with the magnitude of the effect corresponding with the strength of what that signals about the “class” of the rental.

7

8

  • As a host, three easy things you can do to increase the price you can charge: Provide a washer/dryer, Free parking, Elevator in the building, doorman, pool, hot tub.
  • Providing carbon monoxide detector is a plus, but having a smoking detector and fire extinguisher will decrease the price.
  • It is also very interesting to see that heating is a plus, but air conditioning is not. Mainly because everyone things every household will have that, same with hangers and other essential items. As shown in the graph below.

9

Seasonal Effect

An upward trend in pricing was found during major events like ACL and SXSW. As shown in the graph below. Realtors can really up their game during these seasons.

10

Conclusions

  • Rentals downtown, availability of essential amenities, affluent neighborhoods, number of bedrooms/bathrooms will drive a greater demand + price of rentals.
  • Major events such as ACL and SXSW have a large impact on bookings.
  • Even though model’s predictive accuracy is comparatively low (due to outliers in imperfect public data), it definitively points to a large number of interesting facts about the vast opportunity in the Austin market.
  • Hosts must follow the nuanced traveler demands illustrated in order to maximize profit.

 

Sources:

InsideAirbnb, Zillow, Airbnb

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s