Thank you, @TerryT , this is a fun challenge!
I used a similar approach as @kat - it seemed the most straightforward to me by using 'Tokenize' RegEx tool to break the data into rows by character and in the end concatenate the most frequent characters by their position to discern a message. 
