I spotted a couple errors on the Community detection course:
In "Handling missing values" the first Cypher statement includes a line of what should be plain text. Looks like a simple formatting error on the markup
in the KMeans page references to KMeans++ often don't have the ++ characters, so we get things like "The key distinction between KMeans and KMeans".
I think I got that error too, but it was a mistake on my part. I wasn't defaulting missing values to something non-null, so it ended up with vectors of different size. I don't remember if the solution was setting a default value for missing properties or filtering nodes with that property missing from the dataset but once I did that the code ran without issue.