Sneak preview sivis - Turning browser clicks intro reproducible scraping code

For this meetup, we're happy to have our member Tonio Liebr presenting his framework Sivis and giving us insights on web scraping with R.

Sivis enables you to select data in the browser by clicking on it and then generates r code for you to scrape this data periodically.
I am working on this package for over a year now. Its never finished but i guess there is a certain time to put it out there:
Video:
https://www.youtube.com/watch?v=tFZ3os-GoNA
Github:
https://github.com/Toniiiio/sivis

The idea:

When you learn about webscraping you are usually taught to differentiate between javascript and non-javascript pages and use httr/rvest or (r)selenium. You would either send a single request (httr/rvest) or trigger all requests (RSelenium) related to a web page. I think this differentiation is not wrong, but i would propose an alternative view.

"A website consists on average of 70 requests, only one will yield your target information. Why would you want to load the other 69?"

One might argue that it is not easy to identify the correct request among the ~70 candidates and tideous work to collect the relevant meta data for building GET / POST requests. That is correct, if it is attempted to be done manually. But sivis will assist you to automize this process as much as possible.

In my talk we will have a walk-through, some live demo and time for your questions!