This week, we have finally made the transition from paper to software. We have created a few of the previously discussed plots using Vegalite and Vega in R. But first, I would like to talk about the challenges we faced and how we had to make do with what we could in the given time.
The data though in CSV, was extremely huge. We created a combined dataset with 3 cities and that was too large and slowed down the entire process. So, we have decided to remove certain columns that would not be using.
Next, the data columns were not in the formats that we would have preferred. For example, price was a text string with commas and currency signs included, the cities were not written in the same way - some were the French spellings and some were the Dutch spellings, etc. So a majority of our time went into data cleaning and organizing.
Finally, we struggled to recreate our "creative" graphs in the software. Either the shapes and formats were not possible, or the data had too many missing values to be used. We wasted a huge amount of time trying to attain the same graphs rather than work on something new or different.
Nonetheless, we managed to create a few interactive plots and we shall discuss the findings below:
Figure 1: Brussels
Figure 2: Antwerp
Figure 3: Ghent
In the plot above, we were trying to find the neighborhoods that have a requirement of minimum number of nights. If that number is high, then the kind of clientele they accept are those that stay for some work and not necessarily tourists. On the other hand, if it is less than 5, we are guaranteed to have found tourist stay hot-spots. In the bottom, we also have a filter that helps select the city.
It is quite clear from the comparison that Brussels has a wide variety of guests accepted, due to the large range of minimum nights. Brussels also has the highest density of properties in the neighborhood of Central Station, Ixelles and St. Gilles, based on the size of the circle.
Conversely, Antwerp is a little more densely populated than Ghent, having more neighborhoods with bigger circles. Also, Antwerp still has a few neighborhoods that have minimum night as 30 while the range is only till 10 in Ghent.
In this plot we are trying to understand the average cost of renting different type of rooms in the 3 cities.
Hotels in Ghent are the cheaper than the shared rooms
In Antwerp, the best solution is to take an entire house, cause the cost isn't that much higher
DO not take hotels in Brussels, instead opt for a shared room
Finally, please find below the gist links to our vegalite codes: