Altair follows a declarative approach where you specify which columns go to which axis, the type of chart or plot, and what should be interactive. Most tools require you to write detailed boilerplate code to set up the axis and figure.
Ray was originally built for reinforcement learning research, then quietly faded as RL hit a wall. Until ChatGPT showed up. Suddenly reinforcement learning was back, as the post-training step that turns a raw language model into something genuinely useful.