That is the point. When you exclude that group of people only from 15 out of 32 surveyed countries, you skew the results for the whole survey. You can't draw parallel conclusions from different samples.
The goal of this comparison is to compare urban-to-urban, because those countries which don't have this exclusion have relatively tiny rural populations.
That is the conclusion you have drawn, but it's not part of the methodology listed in the survey. They haven't excluded rural participants from the 17 countries, while explicitly excluding them from 15 countries. If you see no issue with that, enjoy your blinders, but please stop spreading misinformation.
Rural populations are negligible and covered under other factors in a number of countries (in the US, Internet access). It's not worth mentioning because it's not a relevant part of data.