This looks interesting but demonstrates how surveys and polls are so unreliable and misleading.
First, I lived in Southern California for over 30 years and everyone I knew called it Coke. I was the odd ball who called it soda/cola and that's because I was tired of hearing the server say "is Pepsi okay?" (I didn't care)
Second, while 120,464 respondents sounds impressive and a good size sampling, that means there are on average around 40 respondents per county. That's hardly a suitable number when you have counties with populations in the 2-3 MILLION range.
So before you go quoting that study, check to see what their data is based on and if you could stand behind those numbers.