jsonlite | A simple , self-contained , serverless , zero-configuration | JSON Processing library

Below, code uses the OP's solution until the conversion to tibble, then we do unnesting of columns individually as the structure is a bit complex i.e. involves matrix as columns in a nested list

You might try this:

Provide a location to write the data in write_json -

They have a common structure by the looks of it. Just extract the t param ,for each link in the link column, and put that into a json url template string:

Do not mix conda install and install.packages. Only use install.packages if the package is not on conda. tidyverse is on conda (see anaconda.org/r/r-tidyverse), so you should now remove.packages("tidyverse") and then conda install -c r r-tidyverse.

You may be also interested in learning about conda-forge channel which provides much more R packages.

If you are also using Python, this also apply to mixing pip install and conda install - doing so will break your installations in the ways that are difficult to imagine (see Is that a bad idea to use conda and pip install on the same environment?). Largely this can be attributed to conda using its own mechanisms for handling library/package paths which and not interacting great with other distribution systems.

The proper way to translate that command to httr is

You can unnest the sessions variable and the slots variable within it to get a long dataframe.

You make headers but never include them in your call to GET. Use them there.

Perhaps the output of jsonlite::toJSON() is ignored because it will enclose your json in []s. What happens if you use rjson::toJSON() instead?