You can't. If you don't know what the population is you can't say anything about it. Statistics is not magic. See: https://www.amazon.com/Elementary-Survey-Sampling-Richard-Scheaffer/dp/0840053614/ref=sr_1_1?keywords=Elementary+survey+sampling&qid=1579477452&s=books&sr=1-1
Sorry, I misstated the problem; the population is known (for example, all office workers in Singapore). But the sample frame is not known, e.g., I don't have a list of all the office workers.
I see, You are correct that is a different issue and the answer depends on how you intend to collect data. I would suggest that you begin with looking at the items at this link. : https://www.google.com/search?q=How+to+collect+data+in+Marketing+Research&rlz=1C1CHBF_enUS874US874&oq=How+to+collect+data+in+Marketing+Research&aqs=chrome..69i57.21689j0j1&sourceid=chrome&ie=UTF-8
I would not recommend a non-probability sample because you really can't trust your analysis. Try to get results you are sure of. If you have more questions, please feel free to ask. Wish you the best, David Booth