Facebook's graph API is quite limited, as Meta prefers to keep the graph APIs for its platforms tightly controlled and dole out access to researchers on a more limited basis. Like Instagram, there's a clause that bans scraping: "You may not access or collect data from our Products using automated means (without our prior permission)." This limits the suite of tools available to collect data from Facebook and makes it so that some of the tools that can do this violate the TOS in the process.
Meta has previously taken action against high-profile research that violates the TOS, shutting down the accounts used for scraping Facebook by the NYU Ad Observatory in 2021. Similar projects investigating ad targeting on Facebook from Mozilla and ProPublica have also been shut down in 2019. Automated web scraping of Facebook is highly cautioned against for this reason.
Meta described CrowdTangle as its "public insights tool", which collects and provides access to public posts on Meta's platforms, Facebook and Instagram (moreso Facebook). Researchers can apply for access to CrowdTangle to get access far beyond what the normal Facebook API provides. This is part of Meta's way of tightly controlling what information it makes available to researchers.
Meta has created hundreds of datasets, mostly country-by-country population mapping and density datasets as well as some datasets of political ads and other topics of concern, that it shares via the Humanitarian Data Exchange repository (HDE). You can find a partial list with descriptions on Meta's Academic Resources page and see the datasets themselves on the HDE.
