What exactly is being shown in the animation/slideshow? You said its iterating and maximizing the MLE, but doesn't that just mean you're maximizing a single number? How can that iteration generate an entire heatmap?
The likelihood itself is a single number, but the algorithm is maximizing the likelihood based on a vector of 993 parameters representing the quality of a seat. Each x,y coordinate represents a seat in the ballpark, and each seat is mapped to one of those 993 values. The heatmap evolves as the vector approaches maximum likelihood.
Cool stuff. Out of curiosity, what are you using to embed Latex in a blog post like that?