Take the following problem. You’re tasked with predicting what a new Airbnb host might charge (and by extension, what a guest might pay) across London. To answer this question you need to collect data about the factors that are likely to contribute to the nightly price of an Airbnb property in the city.

The more of these factors you can collect some data on, the closer your ML model can get to that sweet spot of a near-perfect prediction in the middle.

Produced by Author

You already have information about the 80,000 Airbnb properties in London and their hosts. What you’re missing, however, is crucial information about the competition — where are hotels located, how many shops, bars, restaurants, tourist destinations, etc, are close by. After all, who travels to London to stay in a quiet neighbourhood with no activity? This is where OpenStreetMap can help.

Before we move on, if you would rather follow this guide in the form of a Jupyter Notebook, a link is provided at the end of the article.

#python #data-science #feature-engineering #mapping #openstreetmap