Generalize guesstimate for number of people doing activity at this instant

39 Views Asked by At

From Guesstimation 2.0 question $2.8$:

Q: How many people are talking or texting on their cell phones at this instant?

The key insight to solve a problem like this is that the fraction of time the average person spends doing some activity is equal to the fraction of people doing it right now.
...
$$\frac{t_{talk}}{t_{day}} = \frac{N_{talk}}{N_{day}}$$ ...

I understand this. But this works only when we are talking about the entire world (and the assumption that people doing this activity are uniformly distributed).

If the question is only for a specific region, how would this insight change? For instance, how many people are sleeping in your state at this instant? This fraction of people would be significantly different (more or less) from $\frac{8}{24}$ (assuming an average of $8$ hours/day).