This is a Fermi estimation problem and it's easy.

You have probably heard of similar questions (Google asked them in interviews!) and there is a simple lifehack to solve this by findind an answer within an order of magnitude (10^1):

How many people live in Berlin: 100k, 1M or 10M?

Without googling I would estimate there live more than 100k and less than 10M, so I'll go with 1M.

How many people have pianos in Berlin: 1 out of 10 has one, 1/100 or 1/1000?

Let's go for the middle answer (1/100) since we account for children as well. This means we are estimating 1M people x 0,01 piano distribution = 10k pianos in Berlin

How many pianos can be tuned by one tuner in a year: 10, 100, 1000?

Again, let's go for the middle answer and estimate 100 pianos

Result: 10k pianos / 100 pianos tuned = 100 piano tuners.

I checked Google and Yellow Pages (for real!) and there are around 60 piano tuners in the Berlin metro area.

I would say that's a pretty accurate estimation for the little data we had available!

Startups need to solve problems in constant uncertainty. Estimating an outcome with little data helps navigating the darkness.