Stockholm housing price predictions

In this project we look at the prices of sold apartments in the Stockholm area. The data is scraped from the web and then trimmed and engineered. We then begin by exploring the resulting dataset with some visualisations. Lastly we train regression models in order to predict the price of an apartment given its size, distance from city centre and other variables.

The main goal of this project is not to perform the most accurate prediction, but rather to go through how to get the data from the web with a web scraping tool and how to process the scraped data into something usable.

Contents:

• Data retrieval
  • Web scraping and data trimming
  • Processing the scraped data
• Exploring the data
  • Features
  • Target and predictors
  • Visualising the data
• Predicting the price of an apartment
  • Decision tree
  • Random forest regressor
  • Neural network
  • Gradient boosted tree regressor
  • Summary