Nielsen doesn't do separate measurements in Miami based on language though, does it?
Sort of.
With the database delivered to each subscriber, they can do things like creating a "Black book" or an "Hispanic Book" or even a "Broward Book" as well as full reports by age and gender and income and education. And they can break out "Spanish Dominants" alone, too.
All those groups have specific quotas in each book. So the data, when viewed separately, is just as reliable as any other breakout on one or more of the stratification variables (age, gender, ethnicity, language dominance among Hispanics, county (in multi-county books), education, income.
Remember, there are lots of Black or Hispanic or Spanish Dominant Hispanic buys and they only look at the reports that they can generate for listening by their target group.
I don't know much about the Canadian service. About 22 years ago, I visited the BBM (now Numeris) HQ in Toronto (one of my favorite cities in the world) and discussed language issues. They had one interesting issue which was the measurement of fully bilingual persons who listened to stations in both language. The problem was whether the ratings should be based on asking a person if they were "Francophone" or "English primary" and, beyond that, who was really a Francophone.
In other words, was Joseph Smith whose father managed a lumber mill in Chicoutimi when he was a kid and who grew up speaking almost all French and still prefers it... is a Francophone.