We'll predict future house prices by training a machine learning model to predict if prices will rise or fall. We'll write all of our code in Python using JupyterLab.
We'll use data from the US Federal Reserve, along with house price data from Zillow. We'll merge and combine this data, then use it to train a random forest model. We'll measure error using backtesting, then improve our model with new predictors.
This project can be customized to predict house prices in your metro area if you live in the US.
You can find the full code and README here - https://github.com/dataquestio/project-walkthroughs/tree/master/house_prices .
You can download the data from the previous link, or from this link if it doesn't work - https://drive.google.com/uc?export=download&id=1HlHw_JyRckfPOlwwxUHS-sdDqfZQ732p .
Chapters
00:00 Introduction
02:23 Loading federal reserve data
07:39 Loading Zillow house price data
14:10 Preparing data for machine learning
18:25 Setting up our machine learning target
25:08 Creating a machine learning model
28:11 Creating a backtesting engine
31:58 Measuring error
32:47 Improving our accuracy
35:44 Running diagnostics on our model
40:22 Next steps with the model
---------------------------------
Join 1M+ Dataquest learners today!
Master data skills and change your life.
Sign up for free: https://bit.ly/3O8MDef