I was playing around with the Latitude and Longitude of hospitals in the UK. When I went to try correlation I got NA because of missing data.
cor(hosp$Latitude,hosp$Longitude)  NA
To get round this I created a new variable.
hosp1 <- na.omit(hosp)
This basically omits all NA (missing values). So use the simple function na.omit to create a new dataset without missing data. This excludes any row that has missing data for ANY value.
Then doing this again
cor(hosp1$Latitude,hosp1$Longitude)  -0.1113144
This indicates very weak negative correlation. So as Latitude increase Longitude decreases slightly. But there is no strong connection.