Those questions are nothing more than standard market sizing questions. The reasoning that you need to use is similar to the one you'll use to estimate a market size.
For example to determine how many new-yorkers are wearing a suit at this moment
1) Estimate how many people are in new-york (you can at first consider how many people live in new-york, then consider people commuting to new-york either on a daily basis or for less regular travel). You can split that number into age category (less than 20 y.o, between 20 and 40, etc.)
2) Estimate for each category how many wear a suit and how often (you can for instance use a per weekdays basis). For that you need to find the reason to wear a suit (for a job, for a job interview, for a wedding, for a funeral, etc.) for each segment you need to estimate the frequency of wearing a suit. For instance people below 20 would rarely wear a suit so you can make the approximation that 0% of that age group would wear a suit. Same for people above 60.
3) use your per weekdays number and divide it by 5
At each step you need to do three things:
1) suggest the number you'll use and justify it (for instance i'm in the age group 20-40 and among my friends around 15% wear a suit daily).
2) Ask the interviewer if he thinks this estimation is sound and if he has some more precise data available
3) comment your final number and drive the case to the next step (ok now that i have this number i need to consider that so...)
You'll be tested on your reasoning skills (how do you end up with your result, does it make sense) and less on your final number.
If the question is "at any moment" then you need to go one step further and consider for instance two time periods : between 8 a.am and 8 p.m and outside that time frame.