Is AirBnB secretly dominating Bangkok?
Have you ever wondered how much listing value AirBnB has… in Bangkok?
As you may know, AirBnB is all around us these days. But did you know, how much value it has in itself in a place such as Bangkok, the land of smiles?
This small curiosity has got me wondering and look to identify some insights with data.
Wait….what’s AirBnB?
For those who don’t know or have never heard about this service, they are basically a hotel, but you stay at someone else’s house instead. Checkout their website for more details. https://www.airbnb.com/
So what are we doing here?
As a product manager with technical background, and as a small project, I decided to look into exploring more of the Data Science field and what best learning can one get from doing an actual data project!
Upon looking around, AirBnB has open data where it is freely accessible so I thought it would be nice to do some analysis to find some interesting insights with this dataset. So let’s get started!
Analysis…analysis…analysis!
First off, the structure of this blog is going to take you through a journey of how the data has been analysed and what aspects do we see with this dataset. The approach taken includes:
- View and explore the dataset
- Clean data appropriately
- Analysis and visualisation to answer business questions
Part 1 and 2 can be found in the Git repository here. This blog will focus on answering the following business questions with data.
Overview data
So this is a snippet of what our data looks like for analysis. The data used is about AirBnB’s listings and has details such as their value price, availability, amenities, host id, etc.
Question 1: What are the top 10 most common amenities for AirBnB listings in Bangkok?
This is an interesting that came up in my head. With all these housing and listings, what amenity would be in most places that is offered?
From our analysis, we can clearly see that in Bangkok, AirConditioning seems to be the mostly valued thing people have with up to 18515 listings with it. I guess we cannot live without Airconditioning in Thailand afterall xD
Second to that are things such as Longtermstayallowed and Essentials. This shows that people are looking for a long vacation with a care-free stay as a nice vacation.
Question 2: What are the top 10 areas in Bangkok with the most listing value? How many listings are there?
Here, the next question is exploring how many listings are there and its total value in Bangkok. As shown in the first graph below, there is a lot of variation in different neighbourhood on the listing total value.
Due to this case, it may make it harder to see so we have scope out just the top 10 neighborhood with the highest value listing shown below. As you can see, the neighborhood Vadhana, shows to have the highest total value listing of $5,991,396 having 2441 listings in the area! So if you want to go for a vacation, that’s definitely the place to go to.
Furthermore, we have had a look at the distribution of these listings in each place. As you can see, there are outliers and just one or two listings which have a very high listing value, but otherwise, everywhere else is cluttered in the under $50,000 range.
Question 3: What are the growth trend in number of AirBnB hosts overtime in Bangkok?
Moving on, let’s have a look at the growth trend of AirBnB hosts in Bangkok.
As shown in the table, there has been a big fluctuation in the dates that hosts joined the platform. From the visualised plots, we can see that hosts joins on different days overtime.
From what we can see from the last plot though, is that there seems to be a decline in the accumulated number of hosts of AirBnB in Bangkok from 2020.
Could this be the decline of AirBnB? Or can we blame Covid-19 for this?
Question 4: What is the latest total value of AirBnB listings in Bangkok?
Next up, we want to quickly identify the total listing price and value in Bangkok for AirBnB listings. Coming to no surprise, the total listing amount is around $40 million worth. However, this has not been growing steadily at all as well.
Could this be the deciding factor on the decline of AirBnB in Bangkok?
Question 5: Given an estimated number of listing at the end of the year 2021, what would be our total listing value predictions?
So lastly, since there were a lot of uncertainty regarding whether AirBnB could be in decline in Bangkok, we decide to run some predictions and estimation based on the data that we have.
Data prediction may not be fully accurate as AirBnB only provides data up to 1 year backwards
So based on what we did below, we are looking to predict the sum_price (total listing price) amongst other variables based on the knowledge of date and listing_count.
# Take in date and listing count, we want to predict the sum price and other factors
X_2 = temp_df[['date','listing_count']]
y_2 = temp_df[['sum_price','host_count' ,'superhost_count','has_availability_true_count','sum_number_of_reviews']]# Split data into training and test data, and fit a linear model
X_2_train, X_2_test, y_2_train, y_2_test = train_test_split(X_2, y_2 , test_size=.30, random_state=42)reg = LinearRegression(normalize=True).fit(X_2_train, y_2_train)d_3 = {'date': ['20210822', '20211222'], 'listing_count': [19289, 19289]}
df3 = pd.DataFrame(data=d_3)
y_predict_3 = reg.predict(df3)print('Prediction for 20211222 total value price: \n', y_predict_3[1][0])
print('Percentage diff between 20210321 and 20211222: \n'
, ((40400255.0 - y_predict_3[1][0]) /y_predict_3[1][0]) * 100.0
, '%')
print('Therefore, estimated an increase of 0.5% growth in around 9 months time')
Using basic Linear Regression analysis here, we can see that there is a slight increase of 0.5% of the total listing growth based on our predictions.
So AirBnB is not in a decline after all…or is it…
Conclusion
In conclusion, this small project has shown us the basic insights of the AirBnB business landscape in Bangkok. Analysis methods used are simple calculations and can be found from the following Git repository here.
Thank you for reading and hope you learned something new today!
