Unfortunately I believe it's true that Shiva's analysis is bad. The problem is that by putting the % straight ticket on the x-axis, he's forcing the % of non-straight ticket to go down. This pattern occurs for both Trump and Biden, occurs in other areas, and occurs in previous election years. One goes up, the other necessarily goes down. I'm a data scientist and discovered this on my own while trying to replicate his finding.

This guy does a good job of explaining it slightly different to me: https://www.youtube.com/watch?v=aokNwKx7gM8. Ignore his Benford's "debunking" though, it's ridiculous and unfounded.