Seattle Airbnb data analyzed : what makes a home popular?

Using Seattle Airbnb data to understand trends in popular homes

Seattle (photo credit: me)


A special feature of sharing economy such as Airbnb is that you can be the guest and host at the same time. When you travel around the world staying at someone else’s house, you can at the same time host paying travelers in your home. Or if you have a spare room in your house, you can rent it out to travelers even when you are home. This then-radical idea of sharing your home with strangers has now been widely accepted and used by so many people around the world.

There are a few open datasets from airbnb including the data of my favorite city Seattle. So I decided to explore the data and see what insight I can get from it as a guest as well as a host (I mean future host; have to buy a house/apt first, but anyway).

Part I: Is airbnb rating informative?

What piece of information on the booking website do you trust most when you search for a hotel room to stay? For me, it’s the customer rating. I am the type of person who never goes to a restaurant that has less than three and half stars on yelp and never stays in a hotel whose rating is lower than 7. Don’t say I’m fussy because I believe a lot of you are with me. So I always do plenty of research based on price and rating for every trip, and nearly all my decisions turned out to be good ones. Of course I am making a hasty generalization here, but I do believe that review and rating system provide a lot of information to the customers.

Therefore, my first question on Seattle airbnb dataset is: is airbnb rating system informative just as traditional booking websites?
Distribution of rating scores

To some degree, rating scores of the airbnb confirmed the first impression I got when I browsed its website for my own trip: most of the homes are very highly rated. This can be seen from the distribution of rating scores above as well as the average rating score of 95 out of 100. For me, this is just too good to be true. Well, it is possible that airbnb takes a very good control of the quality of their listings via the rating system, for example by giving warnings to those homes with less than a certain score. If this is the case then, is rating a good indicator of popularity? Because of the sharing feature, privacy and safety have always been concerns for some people (like me), making us naturally trust popular (with more reviews) homes more than others. So, does the rating tell us about popularity?

Relationship between rating and number of reviews

Not necessarily. The ones got really low rating are definitely less popular (less reviewed), but not the other way around. Thus, the take home massage here is that rating alone is not very informative.You would want to combine all the information provided by the host and previous guests, and then choose the one that meets your needs best.

Part II: What features do popular homes share?

Next, I wanted to know what kind of home is popular among all the listings. So I used the monthly review (the average number of reviews a listing got per month) as an indicator of popularity and examined what factors influences the popularity most.

For better visualization, I divided the listings into three groups: homes with low, medium or high popularity based on the number of monthly review.

From the graph here we can see that the most popular homes are the least expensive ones in terms of both total price and cleaning fee (usually is included in the total price).

Price and cleaning fee based on popularity

The minimum number of stay required by hosts is also an important factor as the figure suggests. The average number of stay required by popular homes are somewhere around 2, but it is higher in homes with low popularity.

Minimum number of nights required for stay

Finally, this is a very interesting graph showing relationship between room type, price and popularity. One thing we can see here is that entire home/apt is the most common room type. Among this room type, we see the same trend that we saw above: cheaper homes are more popular than others. However, this trend is not that significant in other room types. So if you are a host of small and affordable entire home/apt, your home will be in great need just by itself (locations are not considered here). But if you own an expensive (possibly huge) house, it’s not a bad idea to rent each room out as private rooms, instead of renting the entire house out.

Part III: As a host, room type and price is not what I can control. What else can I do to improve my home?

Okay, it is true that a lot of things cannot be changed by the host including neighborhood, location or the type of your home. But there are definitely things that we (future) hosts can do to improve the quality of the service thereby increasing the number of visits. I’ll show a few things I found.
  1. Make your home ready for instant booking. Homes allow instant booking are more popular than others.
  2. Respond to potential customers as soon as possible. Most of the popular homes respond within an hour.
  3. Do not make your cancellation policy too strict or too flexible. Most of popular homes have a moderate cancellation policy.
  4. Make sure you provide customers popular amenities. Most popular amenities in Seattle are: wireless internet, heating, kitchen, smoke detector, carbon monoxide detector and fire extinguisher, washer and dryer, shampoo, TV, free parking on premises, family/kid friendly and first aid kit.


In this article, we took a loot at some of the things that we guests/hosts might be interested to know. These include airbnb rating system, features make a home popular, and some minor things that a host can do to improve the service. Some of them may seem very obvious, but it’s always so fun to see how the data confirm our common sense.

To see more about this analysis, see the link to my Github available here.