Our Strategy

  • Showcase code examples with which students can modify and experiment to turn the code into their own projects through learning by doing

Three Barriers to Web Scraping

1. Website that says no

You open a page and it decides you are a robot

2. Pop-up in the way

A cookie banner sits between you and the data you came for

3. PDF you cannot search

Hundreds of reports, and some are just scanned images

We get past all three to unlock the data behind

Scrape sustainability reports at scale

 
Watch the full pipeline run — then take the code home

 

Step 1
Get in         →

past the bot
detection/
cookie banner
gatekeepers

Step 2
Find every       →
report

across a whole
company website

Step 3
Keep the pages
that matter

even when the
PDF is a scan

From texts to patterns

  • How have the risk factors disclosed by the Magnificent 7 changed over a ten-year period?
  • GPU-accelerated natural language processing (NLP) in action:
    • Bigram clouds: word pairs, sized by importance
    • GPU power: see a heavy text job speed up

Why GitHub, GitHub Codespace, and GitHub Copilot?

Meet your AI research partner

Ask

Get answers, explanations and quick help

Plan

Agree on an approach before anything changes

Agent

Let AI write, run and fix code for you

  • Which mode when?
  • Plus API access to a large language model (LLM) for summarising and categorising textual data at scale

If AI can plan and write the code for me, …

  • Why do I need to learn coding?

The more you know, …

  • the better you can stay on top of everything

Otherwise, you won’t notice any wrong direction
AI might take you to — and won’t be able to correct it in time

Story behind Jupyter notebook

  • Embedded in the name Jupyter are letters representing
    • three major data science programming languages:
      • Python, R, and Julia
  • Julia, a high-performance language, is a topic for another day

R

  • a programming language specially designed for statistical computing
    • has become very user-friendly for data analysis and visualization since its tidyverse era

Python

  • a general-purpose programming language widely used for data analysis and machine learning (ML)
    • has become the undisputed industry standard due to an unmatched ecosystem of libraries (aka. packages) that provide pre-built functions for many tasks
  • Pain point of using Python:  
    • Traditional methods to install packages can lead to dependency conflicts and version mismatches

marimo

  • marimo is a reactive Python notebook
    • especially friendly to beginners due to:
      • handling packages installation and environment setup smoothly under the hood

Open-source packages on GitHub

  • Powerful packages are freely available on GitHub
    • The challenge: Buried among numerous poorly maintained/poorly documented packages on GitHub

Adopting Open-source Resources

  • Reliance considerations:
    • Abandoned projects?
    • Organizational backing (e.g., polars)
      vs. Individual developer
    • Active development:
      • Last commit
      • Watch | Custom | Releases, Issues, Security alerts

The month after this workshop

  • Recommendation:
    • Simply keep on integrating AI into your workflow on a daily basis

Getting ready

  • A preliminary self-learning pack gets you ready before the Workshop:
    • Python basics in Jupyter
    • Markdown
    • GitHub Codespaces
  • Nothing to install — everything runs in an updated modern browser

What past participants say

  • “The workshop is beginner-friendly and gives you a chance to try the tools yourself rather than just learning about them in theory.”
  • “Extremely well set up and helpful, would have liked more time on the day or spread over a day and a half.”

English is the new programming language

  • Your productivity is limited only by how clearly you can explain your problem to AI
One day · Hands-on · No prior coding experience needed

See you at the Workshop!