Still a divided city?
Exploration of Airbnb data in Berlin
Introduction
Berlin is the capital of Germany and attracts millions of people from all over the world. Berlin’s popularity continues to grow with 14 million visitors and 34 million overnight stays per year according to VisitBerlin. One of the most attractive and exciting things about Berlin is its eventful and evident history besides its diverse culture, museums, and nightlife.
After the end of the Second World War, the devastated city of Berlin was divided into Soviet and Allied occupation zones. In 1961 the Berlin Wall was built — one of the most iconic symbols of the Cold War. The Berlin Wall runs along the sector border and seals off East and West Berlin for many years. From 1961 until the destruction in 1989 and Germany’s reunification in 1990, both cities grew under entirely different conditions with two separate and distinct city centres. At that time, the neighbourhoods immediately next to the Berlin Wall were unpopular and less expensive. Since 1989 the two former cities and their citizens have grown together again. In some Berlin areas, the former border is still evident in the cityscape; in others, the former scars have already healed.
Airbnb is an online marketplace where people can offer and rent out their houses and apartments. And as in most cities, Airbnb has become a popular choice for people looking for overnight accommodation in Berlin. I want to examine the data of Airbnb Berlin with particular regard to the differences between the former East and West:
- What are the key factors that influence the price?
- Where are the most expensive neighbourhoods?
- Where are the most popular neighbourhoods?
Data
The dataset is made available directly on the Airbnb website . The data used in this analysis are those as of October 13, 2020. Prior to the analysis, I carried out appropriate data cleansing and preparation. The complete source code for this data analysis is provided on Github.
What are the key factors that influence the price?
For the vast majority of people, the price of accommodation plays a decisive role. To this end, we will first examine the key factors that influence the cost of an accommodation offer. The following figure shows the results of correlation analysis concerning the price of accommodation. Features that tend to increase the price are highlighted in green, while features that tend to decrease the price are highlighted in red. Only the five most important positive and negative attributes, respectively, are shown.
Not surprisingly, the analysis shows that the main positive key factors are the number of accommodations and available beds. These are followed by the total number of amenities listed, an available TV set, and high chairs. If it is an entire apartment, the price also tends to be higher. In contrast, the fact that the accommodation has a lock on the bedroom door, a private living room, and a washing machine tends to lower prices. These features are probably more common in hostel-like, small rooms, and shared apartments. Interestingly, if the accommodation is located in the former western part of Berlin, the price tends to be lower. The negative correlation of the longitude supports this fact. However, no particular district, neither from the former West nor East Berlin, seems to be correlated well with price. However, it must be pointed out that except for accommodations and beds, all attributes have only a poor correlation with less than ±0.3.
Where are the most expensive neighbourhoods?
The observed weak negative correlation of the western districts with the price could indicate that the former western part of Berlin is less expensive than the former east. But are we also able to distinguish visually between the two former cities? Lets display the median prices in each neighbourhood and highlight the former Berlin Wall to separate the two parts. To narrow down the analysis, we will only consider accommodations for two people with one bed.
The map shows that the most expensive districts are located in the central part of Berlin. Some of Berlin’s peripheral districts also show relatively high prices, probably due to their beautiful nature in the immediate vicinity. The most expensive central districts are located in the former eastern part of Berlin, namely in the eastern parts of Mitte and Prenzlauer Berg. These two districts are considered hip, modern, and popular, which explains the high prices. While these two highly expensive eastern districts probably affect the average prices between East and West Berlin, there is visually no other clear distinction between East and West.
Where are the most popular neighbourhoods?
Finally, we want to investigate which are the most popular neighbourhoods. The question is how we define a popular neighbourhood. For this brief analysis, I naively assume that the number of reviews represents a popular neighbourhood somehow. More precisely, we will analyse the number of reviews per number of listings in each neighbourhood. Why not check whether a review score is high or low? The reason is that the vast majority of reviews in the dataset are positive. The next figure shows the sum of the reviews per listings in each district and the course between East and West Berlin.
The heat map clearly shows that the most popular districts are located in the central part of Berlin, around the former Berlin Wall. The peripheral districts seem to be much less popular — i.e., to have fewer ratings per listings. There is no clear distinction between the former eastern and western parts of the city. The assumption made here to define a popular neighbourhood is vague but should prove useful insight.
Summary
Based on this brief Airbnb data analysis, it seems that Berlin has grown together again, with expensive and popular neighbourhoods in the central part of Berlin — like most other cities. The primary key factors are the number of accommodations, beds, amenities, and the type of housing, say a whole apartment or just one room — no surprises here. The most expensive districts are centrally located in Berlin, e.g., in the former eastern part of Mitte and Prenzlauer Berg.
Like most other cities, Berlin has rebuilt its unified city center, with its advantages and disadvantages. The times of low-priced and heterogeneous inner-city districts are over, as not only the Airbnb data but also the rapidly growing housing market shows.
