Build an App to Find Underpriced Apartments
In Chapter 1, The Python Machine Learning Ecosystem, we learned the essentials for working with data. We'll now apply that knowledge to build out our first machine learning application. We'll begin with a minimal, but highly-practical example: building an application to identify underpriced apartments.
If you've ever searched for an apartment, you will appreciate just how frustrating the process can be. Not only is it time-consuming, but even when you do find an apartment you like, how do you know whether it's the right one?
Most likely, you have a target budget and a target location. But, if you are anything like me, you are also willing to make a few trade-offs. For example, I live in New York City, and being near an amenity like the subway is a big plus. But how much is that worth? Should I trade being in a building with an elevator for being closer to the train? How many minutes of walking to the train is worth walking up a flight of stairs? When renting, there are dozens of questions like this to consider. So how can we use machine learning to help us make these types of decisions?
We'll spend the remainder of this chapter exploring just that. We won't be able to get answers to all the questions we have (for reasons that will become clear later), but by the end of the chapter, we'll have created an application that will make finding the right apartment just a little bit easier.
Here's what we'll cover in this chapter:
- Sourcing apartment listing data
- Inspecting and preparing the data
- Visualizing the data
- Regression modeling
- Forecasting