What is the secret sauce of great listings on Airbnb?

A machine learning-based deep dive into Seattle to uncover insights

The analysis is comprised of three parts, where we start by addressing a few fundamental questions of Seattle short term rental market by understanding time and price, location. We then investigated the drivers of great listing in Seattle.

To start with, one of the first questions that is very crucial for me personally, as an avid travel enthusiast, and for current and potential Airbnb hosts is

We reviewed the dataset which is from the year 2016 to beginning of 2017. To address this question, we developed a metric which is the availability/unavailability ratio. The ratio is defined as the total number of available listing divided by the total number of unavailable listing on a single day. By looking at this metric over time along with the average listing price, we observed:

It appears that the Summer, in particular, June to Aug is the most popular time to visit Seattle. I don’t think it would come as a surprise to anyone, given everyone is enjoying the sun, the water and everything else. Plus, a little less rainfall is helping too, which would also explain that Seattle is becoming less popular in fall and winter when there is more rainfall.

We also confirmed the total number of listing remained pretty much flat throughout the year, which affirms the trend is not caused by an increase in Airbnb listings.

We also noted that the average price increased about 10% during that time.

Next, as everything in real estate, location, location, location. Thinking about what different neighborhoods in Seattle offers, let’s do some data mining to answer this:

We performed a Wordcloud analysis by extracting the keywords from the descriptions and here is what we found on the main areas of Seattle. Let’s take a look.

Finally, the main event: what is the serect behind a great Airbnb listing? In order to answer this, we will analyze what is influencing and driving the review scores provided by customers. Let’s answer:

We started with data cleaning and some exploratory analysis to prepare the data for modeling. We used Gradient Boosting algorithm and through feature engineering and hyper-parameter tuning, we achieved a Mean squared error of about 33. The below are the main findings based on the feature importance analysis:

This gives us some hints about how to become a great host in Airbnb. While this suggests it will take time and experience, as everything in life, this also serves as a reminder to us that it is ultimately about how to delight the customers, build trust with them, and make their experience truly wonderful.

Related posts:

A bicycle ride through a crowdy promenade can be nerve-wracking. Depending on your soft skills you can improve and learn a lot. If only you recognize the opportunity.

I hate to say it but government taxation is a necessary evil that is essential to your business. Paying your taxes is now thought of as one of the most heinous acts a person must do in their life…

How to succeed in business? Every entrepreneur should be aware of this simple idea to climb the ladder of success