3.2. Collecting data, web scraping with curl

Linux for Data Scientists
HOGENT toegepaste informatica

Thomas Parmentier, Andy Van Maele, Bert Van Vreckem, Jan Willem

2025-2026

Interacting with web servers from the terminal

Terminal-based web clients

Browsers:

CLI utilities:

cURL

$ curl icanhazip.com
193.190.172.117

Try this:

  • curl 'https://icanhazdadjoke.com/'
  • curl 'https://api.coinlore.net/api/ticker/?id=90'
  • curl 'https://education.thingsflow.eu/IAQ/DeviceByQR?hashedname=5201731f632701e602d31f98be7297e088a94eb38736c452495f02e444d4ba2d'

Output, redirection

  • Normally, curl prints to stdout

  • When redirected, progress information is printed:

    $ curl 'https://icanhazdadjoke.com/' > joke.txt
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100    95  100    95    0     0    320      0 --:--:-- --:--:-- --:--:--   319

  • Turn off with -s / --silent

HTTP Request methods

E.g.

curl -X GET https://httpbin.org/anything
curl -X POST https://httpbin.org/anything
curl -X PUT https://httpbin.org/anything
curl -X DELETE https://httpbin.org/anything
...

Which request is default?

Saving the result

  • -o, --output <file>
  • -O, --remote-name
curl -s -o anything.json https://httpbin.org/anything
curl -s -O https://www.google.com/robots.txt

Show headers

  • -i, --include: toon ook de HTTP response headers
  • -I, --head: toon enkel de headers
curl -i http://google.com

Set request headers

curl -H 'User-Agent: Mozilla/4.0 (compatible; MSIE 5.0; Windows 98; DigExt)' \
    https://httpbin.org/anything

Follow redirects

  • -L, --location
curl https://www.twitter.com      # Leeg resultaat!
curl -i https://www.twitter.com   # zie location: veld in de response header
curl -L https://www.twitter.com   # volg de redirect

POST data

  • -d, --data var=val&var=val
curl -X POST -d 'penguin=tux&color=blue' https://httpbin.org/anything

Opdracht Automatiseren data-workflow

  1. Verzamel data over een bepaalde periode (bv. curl)
  2. Zet ruwe data om in geschikte vorm (bv. JSON/HTML -> CSV)
  3. Simuleer analyse van de data (bv. grafiekje, basis-statistieken)
  4. Genereer rapport (webpagina, PDF)

Resultaat in te dienen/demonstreren (= 30% examencijfer)

Stap 1. Verzamel data

  • Kies een dataset
  • Schrijf een script dat de gewenste data downloadt
    • Slaat op in bepaalde directory (instellen met variabele)
    • Bestand met timestamp in de naam

Voorbeelden open datasets