The impact of price on reviews for the 100 most expensive and 100 least expensive private rooms of Airbnb

"To what extent do the 100 most expensive private rooms get more yearly reviews than the 100 cheapest private rooms compared across 5 Western-European capitals (Amsterdam, Berlin, Brussels, London, Paris) on an annual time base? An analysis of multiple cities and time-frames."

Project Motivation

Leaving a review and sharing it with the world wide web is only a matter of seconds. The rise of the internet gave room for leaving reviews about anything, anywhere, very easily. Whereas a few decades ago people relied more on word of mouth and written reviews in papers, nowadays we have access to millions of reviews in just a few mouse clicks. For many people it is important to check the reviews before deciding to book the room.

Specifically, we are curious whether the amount of reviews is higher for the top 100 most expensive private room compared to the bottom top 100 least expensive private rooms. The expectation is that individuals who book more expensive rooms prioritize a higher number of reviews as a sign of quality assurance. However, people who stay at budget Airbnb's might be more critical of booking a place with more reviews to ensure it is a decent yet cheap place. In order to gain a deeper understanding of the amount of yearly reviews of private rooms on the most expensive and least expensive Airbnb's we will conduct a research. Due to the unavailability of extensive data, we will conduct the research using a limited datset, which may not meet our desired data volume. Hence, we will compare 2 years worth of reviews and analyze the findings to answer the research question described above.

Our study investigates the relationship between the price and popularity of private rooms in five major Western European cities, offering insights into the dynamics of the tourism industry. Comparing the annual reviews of the 100 most expensive and 100 cheapest rooms across different timeframes provides valuable information about the impact of price on accommodation demand in various cities and years. This can assist policymakers and entrepreneurs in making decisions related to pricing strategies and market analysis.

Method and Results

Method

The importance of considering the pricing on reviews is to gain a better understanding of the relationship between the yearly reviews and the pricing of private rooms. The research method chosen is an analysis of the ouput of the available data in 5 citites in 2022 and 2023 on Airbnb. The 2 categories are coded as the 100 cheapest private rooms per city (e.g. AMS_cheap) and the 100 most expensive private rooms per city (e.g. AMS_expensive).

In this research, we have selected 5 capitals from western European countries. The reason for this is that those cities are most likely to have similar circumstances and thus will provide a solid background to research the question. The five selected cities are:

• Amsterdam
• Berlin
• Brussels
• London
• Paris

Two data points are used. The top 100 most expensive and the 100 cheapest private rooms will be further examined in the research.

Variable types

The following variable types are used in our research:

| Variable     | Description                       | Data class     
| -------------| ----------------------------------| ------------
| City         | City of observation               | character  
| Price        | Listing price of room per night   | numeric  
| Date         | June 2023 & December 2022         | date  
| Room type    | Private room                      | character  

Type of analysis

To adress our research question, we first decided to take a close look at the output of our code. After which we will create barcharts of the chosen variables. The dependent variable in this research is the amount of reviews on average that cheap or expensive rooms in a city have received and the independent variable is the price of these private rooms. We decided on adding 5 different cities to see if the location might also have an effect. By adding a control variable, the aim is to get a better understanding of the effect of yearly reviews and whether this finding is consistent over multiple cities.

Results

After running our code, we figured out that the results are slightly inconsistent and thus it is harder to generalize a conclusion of the data.

By visualizing the data into barcharts we got the following insights. In Amsterdam the cheap private rooms consitently had more yearly reviews than the expensive rooms. On the other hand, Berlin, Brussels and London showed opposite results as the expensive private rooms consistently got more reviews. For the city of Paris the results were surprisingly different, as both the cheap private rooms and the expensive private rooms got almost an equal amount of reviews. Therefore, the generalization of the data would be that expensive private rooms would recieve more yearly reviews. However, this finding is only consistent for 3 out of the 5 cities and thus further research needs to be conducted to gain a deeper understanding of the effects of the amount of yearly reviews and the pricing of private rooms.

Repository Overview

  ├── src
   ├── analysis
      ├── analyze.R
      ├── makefile
   ├── data-preparation
      ├── clean_data.R
      ├── download_merge_data.R
      ├── makefile
   ├── paper
      ├── descriptives.Rmd
      ├── makefile
  ├── .gitignore
  ├── README.md
  ├── makefile

Acces the output by ../gen/paper/output/descriptives.html

Dependencies

For a smooth execution of the project, follow the installation guides on Tilburg Science Hub

• R
  • Setting up R and R studio
• Make
  • Setting up Make
• Github
  • Setting up Git and Github
• PanDoc To knit RMarkdown documents, make sure you have installed Pandoc using the installation guide on their website (https://pandoc.org/installing.html)

If you run the whole script all necessary packages will be installed. Alternatively, you can install the following packages here:

install.packages("readr")
install.packages("tidyverse")
install.packages("dplyr")
install.packages("ggplot2")
install.packages("shiny")
install.packages("png")

Example of workflow research reproduction

• Running the code by using Make

  1. Fork this repository
  2. Open your terminal and run the following code git clone https://github.com/{your username}/airbnb-effect-of-price-on-reviews.git
  3. Set you working directory to airbnb-effect-of-price-on-reviews_ and run the following command: make
  4. Make will run all the code and will generate a html file with the barchart results of the analysis
  5. Ensure that the pipeline is cleaned of all unnecessary data files, running the following code in the terminal: make clean

• An alternative route to run the project using the repository context

  1. ../src/data-preparation -> download_merge_data.R
  2. ../src/data-preparation -> clean_data.R
  3. ../src/analysis -> analyze.R
  4. ../src/paper -> descriptives.Rmd
  5. Open output ../gen/paper/output/descriptives.html

Resources

• Insideairbnb

Authors

This repository is developed by team 10 for the course Skills: Data Preperation & Workflow Management (Fall) under supervision of Hannes Datta. This course belongs to the marketing program Marketing Analytics of Tilburg School of Economics and Management. The team consists of the following members:

• Max Baaten
• Denise Baltus
• Rick Massuger
• Sjuul Visschers