Visualization assignments using Shiny at the course of Big Data on the EIT Digital data science master at UPM
The rise of vacation renting and platforms like Airbnb has created a lot of issues in big cities, like Madrid. We have decided to analyze a public dataset which offers information about Airbnb renting, dates and zones. The dataset is public in Kaggle: Madrid Airbnb dataset.
This dataset is well structured, giving the opportunity of extracting a lot of insight from it, also it has a considerable volume of data. We chose it because of its great potential and the various visualizations that could be extracted from it.
The objective of this work is to answer several questions according to the presented problem domain using the visualization tool Shiny. The questions we are trying to answer are:
- How does the renting price change depending on what neighbourhood of Madrid the place is? Does it vary too depending on the date?
- What are the neighbourhoods with the best ratings? (taking into account all the categories of the scores)
- What are the neighbourhoods with the most number of Airbnb flats? Has this changed over time?
- What insights can we gather from the review comments? What do clients value more from the Airbnb?
The project was developed using R programming language, Shiny framework and the library shinymaterial to improve the visual appareance of the input design.
The full report can be found at project/report.pdf and it contains the visual approach taken to tackle the research questions, along with a guided use of the application. The interface is publicly available on the website https://angeligareta.shinyapps.io/shinyairbnb/ (it can take a moment to load as it is hosted in shinyapps).
After developing this tool and make and exhaustive study of the dataset we arrived at some interesting conclusions.
First of all, speaking about the price, we can see that the most expensive neighbourhoods are those in the residential and fancy part of Madrid (like Salamanca and Goya), with some exceptions due to high rent places (outliers) located in other parts of the city.
To see if this is due to a problem with the offer and demand in that zone, we can use the other visualization to see the amount of places available in that zone. The number is really small compared to other parts of the city like the centre, but similar to other “extra radius” listings, which would validate our assumption that the zone is overall more expensive.
Considering the score question, we can see that mostly in all the categories, all the zones have a good punctuation. The only exception is when we are considering the location punctuation, because the most central places are the ones which have a highest punctuation (Sol, Recoletos, Justicia...).
Finally, regarding the number of flats, we are going to compare for example the number of flats that were in Sol in the year 2012 and in the year 2019. As we can see, in 2012 there were only two flats in Sol, but in 2019 there are 421. So we can conclude that each year more and more flats are added to the platform.
- Cristian Abrante - CristianAbrante
- Guillermo Antoñanzas - Buhorl
- Angel Igareta - angel@igareta.com