The great majority of the projects about machine learning or data analysis I write about here on Bigish-Data have an initial step of scraping data from websites. And since I get a bunch of contact emails asking me to give them either the data I’ve scraped myself, or help with getting the code to work for themselves. Because of that, I figured I should write something here about the process of web scraping!

There are plenty of other things to talk about when scraping, such as specifics on how to grab the data from a particular site, which Python libraries to use and how to use them, how to write code that would scrape the data in a daily job, where exactly to look as to how to get the data from random sites, etc. But since there are tons of other specific tutorials online, I’m going to talk about overall thoughts on how to scrape. There are three parts of this post – How to grab the data, how to save the data, and how to be nice.

As is the case with everything, programming-wise, if you’re looking to learn scraping, you can’t just read tutorials and think to yourself that you know how to program. Pick a project, practice grabbing the data, and then write a blog post about what you learned.

There definitely are tons of different thoughts on scraping, but these are the ones that I’ve learned from doing it a while. If you have questions, comments, and want to call me out, feel free to comment, or get in contact!

Grabbing the Data

The first step for scraping data from websites is to figure out where the sites keep their data, and what method they use to display the data on the browser. For this part of your project, I’ll suggest writing in a file named gather.py which should performs all these tasks.

Continue reading →