I am an avid user of Pocket for saving articles that I want to read later. It’s great, but I have a long queue with anything between short news and 40min+ longreads. This makes it a bit difficult to find articles that I want to read now, as time is part of the consideration beside the article itself.
In the past, I had used ReadRuler to generate tags in Pocket based on my own reading speed: I could very simply filter on articles that take me approximately 2 minutes to read when I’m on the bus and need to get off soon. I can find a great longread when I am slowly waking up on a Sunday morning.
However, I recently learnt that ReadRuler will soon shut down: Chris Pederick, who had created and maintained ReadRuler, recently announced that it will be shut down on May 1st (see the announcements here and here).
This is sad news, but I quickly decided to scratch my own itch and develop a simple R script that does the same job: Taking all untagged articles and classifying them based on my own reading speed. I’m leveraging the pocketapi package from CRAN (where I contributed a few lines of code in 2020) and some base-R functions to have the same function.
Below is the gist which I published on Github. If you want to use it, make sure that you have set up an app in the Pocket API and stored the credentials for authentication in your RENV. This procedure is explained in the link at the top of the gist.