It's not the quantification that makes it important, it's that it's being put into context. When you see a list of twenty tweets holding the same viewpoint, the main question is whether they are representative (and of whom are they representative). Whether you say "70%" or just "most people," it's the same value that's missing usually. The quantification doesn't matter.
It's the difference between a music critic choosing the 100 best songs (fine, because it's a person's own opinion presented honestly) versus a music critic saying they're just going with the popular consensus while cherry picking tweets that happen to pick the 100 they would have chosen (bad, because the tweets aren't put into context). Or like a traditional paper looking for a scientist who will say something exciting/sensational in a particular direction, then presenting it as scientific consensus.
You can find a group saying anything on twitter. That means that the fact that some particular thing exists on it is useless.
It's the difference between a music critic choosing the 100 best songs (fine, because it's a person's own opinion presented honestly) versus a music critic saying they're just going with the popular consensus while cherry picking tweets that happen to pick the 100 they would have chosen (bad, because the tweets aren't put into context). Or like a traditional paper looking for a scientist who will say something exciting/sensational in a particular direction, then presenting it as scientific consensus.
You can find a group saying anything on twitter. That means that the fact that some particular thing exists on it is useless.