Don't use NaNs in tables when the value simply hasn't been recorded

[jimc101]

In several instances, NaNs are recorded in statistical tables. This is often simply because the value wasn't calculated for some reason (e.g., the was no information for a given region).

In these instance, I recommend not using NaNs. Simply put a "-" or similar to indicate the value wasn't calculated. NaNs should be reserved for instances where a NaN was calculated in a given mathematical operation (e.g., division by zero), as this will assist debugging.