In the second week, we tried to explore deeply of our final dataset and drew the basic structure of research questions. Figure 1 was the slides we have presented and discussed with Professor Jan Aerts. We got some valuable advice and wanted to appreciate him again.
Figure 1: Analysis of AirBnB Data
Firstly, the scope of the dataset was limited in Belgium since it can help us concentrate on the techniques and prevent us from getting lost. Secondly, we decided to use the bar chart and scatter plot to depict some basic features of the dataset. It can summarize basic information clearly and quickly. Thirdly, the structure of the research question was built and also presented in Figure 1. We may try to visualize the data in three directions:
the property information
the review behavior of users
host management
This is because we assume that the users’ reviews can be affected by the information about room property and host management. From a descriptive view, the information is also meaningful and it can be used to give practical information for readers, like listing the top 10 popular rooms in some cities of Belgium.
In detail, the property information contains building type, amenities availability, and location (neighborhood) information. The users’ review behavior includes the factor ratings, the number of reviews, and the average ratings per month for each place. The hose management incorporates information about how long the person has been a host, how fast the host replies, the demographic information, acceptance rate, and cancellation policy.
After identifying the scope and research question of the dataset, we also discussed the distribution of tasks and planned to draw sketches in the next week before the formal coding by vega.