Once we had the main question in place, we wanted to start digging the data - trying to find some patterns or some interesting findings. We started with a basic exploratory analysis - comparing several variables with each other.
Figure 1: Number of listings per Neighborhood in Amsterdam
In the above figure we wanted to see of all neighborhoods have the same number of listings or if some were more popular. But it is clearly visible that certain neighborhoods like Oud-West and Oud-Zuid have more listings. This could be due to those areas being the central hub or it could be that real estate prices were lower there.
Figure 2: Avg Reviews per Neighborhood
This graph was specially interesting to see as the results are in contrast to the previous one. It was expected that the neighborhoods with highest number of listings would be popular and would also have the highest number of people posting reviews on a monthly basis. But this graph shows that it is not true. De Wallen and Nieuwendammerham have the highest number of reviews per month. This could be due to the fact that these locations have "cool" and "hip" or up and coming listings that are doing very well and get continuous trail of tourists.
Figure 3: Host Response Time vs. Total Reviews
It is quite intuitive that tourists are more likely to leave a review for someone who was extremely courteous, friendly and maybe helpful. So, if the host was to reply to all their queries quickly, we can assume that they are more likely to leave reviews. This plot indicates the same. Hosts that take more than a few days to respond have significantly lower number of reviews than those who take a day, which in turn is lower than those who take a few hours.
Figure 4: Square-foot Vs Price
We had initially imagined that a bigger place would cost higher and a smaller place should cost lower. But this graph indicates otherwise. There are 3 unique groups here:
Top Circle: These are small properties but are expensive - could be some Boutique rooms
Bottom Circle: These are large rooms which are pretty cheap - could be shared dormitory style rooms
Following the hypothesis (rest): These property prices linearly increase with size
The following posts would try some creative ways to find more stories that this data presents