In this model there will be two equations, one with different types of waterfront properties will be examined. The other there will just have a dummy for waterfront or no waterfront. Every independent variable will not be explained, as some choices are obvious. However interesting variables, changes, and drops will be mentioned. The process for creating this model will be explained in the following few pages, in roughly the order everything happened to show the thought process.
Instead of using months, seasons are used. The market is often active in the summer as parents prefer to move while their kids are out of school. There is a dummy variable created for spring, summer, and fall. Year sold is an important variable to allow for general fluctuations in the housing market and the economy as a whole. Years was decided against at first because it looked like a sin function, however dummies for each year was added later on.
Created waterfront dummy variable first. One interesting thing that was found out, while creating this dummy variable that there are exactly 50 houses on each of the four waterfront types for river, canal, bay, and lake. This is either a coincidence or it is possible whoever created this data set cherry picked it so there was an equal number of each water type. The only one that was different was ponds that has 397 houses on ponds.

For the regression that looks at a more detailed version of waterfronts.
For River
Create riverdum, for being on a river, but not on a premium river. Create dummies for the Alafia and Palm rivers as well, since they are the premium rivers. However after doing this, there is no houses on the palm river so it is dropped.
For Canals
Since sailboat access might have an extra premium create dummy for Bay canal and then one for being on another canal. One thing to note is that there are only four homes