On visual comparison between the 97th Congress and the 116 Congress, it's obvious that the Democrats have moved pretty far downwards as well as bit leftwards. That is, they became much more liberal on the "Other Votes" dimension while moving much less on the "Economic/Redistributive" axis. Interestingly, this leaves Democrats and Republicans with a pretty similar distribution along that axis.
I don't think this is a good method for comparing different time periods. In the 36th and 37th Congresses, the diehard pro-slavery Democrats versus the abolitionist Republicans, the graphs put the parties in the same relative positions as today and make the parties look less divided than today, even though the Democrats felt like they could no longer co-exist and seceded from the Union. It should also be obvious that the slave-economy Democratic Party was vastly different than the modern-day party, but the graphs don't reflect that.
Use a source like the extremely well-researched Dark Money and it'll be evident that the GOP fundamentally isn't representative of the people, it doesn't even pretend to be, and if you listen in the right forums, their leaders don't even claim that. Yet, people, trained in data driven thinking or not, are boxed in into thinking with their guts, how they 'identify', and are taken for a run.
Maybe the strategy of making you question the very possibility of even knowing truths about political actors is working as well as the instigators of such theories think they would.
https://voteview.com/congress/senate https://en.wikipedia.org/wiki/NOMINATE_(scaling_method)
https://voteview.com/congress/senate