Gee, I got it out of the Wall Street Journal, a newspaper owned by Rupert Murdoch (you know, Mr. Fair and Balanced himself). If you'd read the article, you'd know that. But reading is a strain for wing-nuts, which is why Rush has a career and the money to support his drug habit and his appetite for steaks.
You provided a link to a story, not a link to the actual data. Of course, there was no description of methodology and I am left with the concern that the reporter made their own conclusions from the data.
Again, the example of California Pizza Kitchen comes up: while people are not registered as "liberal" and "conservative" the percentage of registered Democrats and Republicans in the areas of California where the chain operates quite precisely mirrors the percentage of "liberal" and "conservative" consumers listed in the table.
The more logical conclusion is that restaurant user mirror to the greater extent the population residing in the immediate area. A chain with more stores in conservative areas will have, thus, more conservative consumers.
Remember that Experian mines the same "highly accurate" (deep sarcasm intented) data they use to produce credit reports to extract this "survey data".