The present article summarizes in a non technical way the the web app solution I created. To dive deeper, I kindly invite you to visit my Github repository to read the underlying R code.
Stack: R / Rstudio, Rshiny
Tourism is one of the biggest sectors of the French economy. Since the early 2010s, the number of foreign tourists has almost never stopped increasing (from 83.6 million in 2013 to 90 million in 2020, source : DGE - Banque de France). In this context, what importance does a pure player like Airbnb have in France? Does the French website host many active listings or is it still rather confidential?
Using some Insideairbnb.com data, I constructed a web app that aims at summarizing Airbnb listings in 3 major French cities -Paris, Bordeaux, Lyon - as of end October 2022.
Airbnb does not provide country specific statistics in its SEC annual fillings (it gives at most figures for the EMEA market). Therefore, to make a case study about the French market, I had 2 options:
This is the latter solution that I chose. I relied on the data provided by Insideairbnb.com website, a project that provides data and adovcacy about Airbnb's impact on residential communities. The project collaborators have already analyzed, cleansed and aggregated the data to facilitate public discussion.
The data on Insideairbnb.com are distinct for each city. Thus, to make a comprehensive study about the French market (global and city-comparative visions), I had to download and concatenate the detailed data and then perform my own analyses.
Screenshot of the Insideairbnb website (Paris focused)
My goal in this project was to construct a web app that gives a broad overview of the French Airbnb market as of end of 2022. More specifically, thanks to the new dashboard, I wanted to answer the following questions:
Data import
Data cleaning:
Illustration of the R code to clean the format of important variables
My R code is separated into 2 files:
"Airbnb_Rcode_French_market.R". This R code prepares all the data necessary for the Rshiny app to work. For instance, it imports the raw data, cleans them, computes summary statistics, prepares the data for the wordcloud, performs the hosts kmeans segmentation, etc.
"Airbnb_Rshiny_app.R" : The Rshiny app structure. The file contains both the server and the UI parts. In the server section, I called all the objects I wanted to display in the app. The UI section on the other hand specifies how the information must be structured in the app (e.g. How many tabs ? Which color ? How many columns ? etc). This code was used to deploy the app on the website shinyapps.io.
As for the creation process, I used frequently a pen and a sheet of paper to draw the design of the app I wanted to achieve.
The landing page gives an overview of the French Airbnb market thanks to colourful BANs.
The "Listings" tab aims at providing statistics about:
The "Pricing" tab aims at providing statistics about:
The "Hosts segmentation" presents the results of the kmeans clustering:
The "Review score" presents the correlations between the guests satisfaction scores and some listings and hosts characteristics
The "Maps" tab represents all listings for each city in a map. The darker the colour the more expensive. The larger the circle the more reviews the listing has received.
Guests seem pretty happy with Airbnb listings (mean review score of 4.6 / 5). Higher customer satisfaction is often correlated with the amenities available in the accommodation (the more, the better) and the host profile (superhost that has already received several reviews).
Hosts do not have the same profile and can be categorized into 5 segments: The Professionals, the One-shot / Lost, the Early Adopters, the Ambassadors and the New hosts.
Send me a message
