README.md

IRE 2024: Web scraping with Python

🔗 bit.ly/ire24-python-scraping

This repo contains materials for a half-day workshop at the IRE 2024 conference in Anaheim on using Python to scrape data from websites.

The session is scheduled for Sunday, June 23, from 9 a.m. - 12:30 p.m. in room Orange County Ballroom 3.

First step

Open the cmd application. Copy and paste this text and hit enter:

cd Desktop\hands_on_classes\20240623-sunday-web-scraping-with-python-pre-registered-attendees-only && .\env\Scripts\activate && jupyter lab

Course outline

• Do you really need to scrape this?
• Process overview:
  • Fetch, parse, write data to file
  • Some best practices
    • Make sure you feel OK about whether your scraping project is (legally, ethically, etc.) allowable
    • Don't DDOS your target server
    • When feasible, save copies of pages locally, then scrape from those files
    • Rotate user-agent strings and other headers if necessary to avoid bot detection
• Using your favorite brower's inspection tools to deconstruct the target page(s)
  • See if the data is delivered via undocumented API to the page in a ready-to-use format, such as JSON (example 1, example 2) -- Postman or similar software is handy for testing out API calls
  • Is the HTML part of the actual page structure, or is it built on the fly when the page loads? (example)
  • Can you open the URL directly in an incognito window and get to the same content, or does the page require a specific state to deliver the content (via search navigation, etc.)? (example)
  • Are there URL query parameters that you can tweak to get different results? (example)
• Choose tools that the most sense for your target page(s) -- a few popular options:
  • requests and BeautifulSoup
  • playwright (optionally using BeautifulSoup for the HTML parsing)
  • scrapy for larger spidering/crawling tasks
• Overview of our Python setup today
  • Activating the virtual environment
  • Jupyter notebooks
  • Running .py files from the command line
• Projects in this repo:
  • South Dakota WARN notices
  • U.S. Senate press gallery
  • IRE board members
  • Queensland workplace incidents
  • Texas custodial death reports

Other resources

• Try GitHub Actions if you need to put your scraper on a timer (you could also drop your script on a remote server, such as DigitalOcean, PythonAnywhere or Heroku, with a crontab configuration
• Tipsheet on inspecting web elements
• Tipsheet on saving HTML files before scraping them
• Tipsheet with some miscellaneous scraping tips

Running this code at home

• Install Python, if you haven't already (here's our guide)
• Clone or download this repo
• cd into the repo directory and install the requirements, preferably into a virtual environment using your tooling of choice: pip install -r requirements.txt
• playwright install
• jupyter lab to launch the notebook server