Correlation between numeric variables of different meaning I have a data set containing information about buildings, such as the area code, the neighborhood code, the number of floors, the year in which the house was built and the year in which restorations were made. All of them are integers.
I want to see if any of these variables are strongly correlated to the number of incidents reported in the house (such as problems with the heating, fire alarms etc.). I looked at the correlation matrix and there seems to be almost no correlation between the number of incidents reported and the data.
I'm wondering whether I'm applying the wrong method and whether there is another way of finding a correlation between the two. How do I make a distinction between, say, a 4-digit number that represents the ZIP code and the year in which the house was built? 
 A: A zip code would usually be treated as categorical, since there is (presumably) no meaning to the actual value and difference between the numbers, or ordering. 
The year of building would usually be numeric, since there is a meaning to the numbers themselves - 1990 is earlier than 1999, and (as I write this in 2020) a house built in 2010 is twice as old as one built in 2000 for example (you can't say the same thing about zip codes and other categorical data).
A standard approach would be a regression model, where the number of incidents is the outcome (response) and the other variables are predictors (explanatory variables). If the data are clustered within zip codes then you might want to use a mixed effects regression model with random intercepts for zip code (and any other categorical variables that are grouping variables (such a city or neighborhood). The model will produce estimates for the explanatory variables being the association of each one with the outcome, while holding the other constant.
