Mostly, I think that any tool of this sort that somehow becomes sufficiently popular and effective to actually make a measurable difference to US presidential election results, relative to the effects of advertising and blogs and newspaper articles and so forth, will be co-opted by a raft of deliberately biased competitors long before that point.

That said, I do think something similar would be useful for local elections. Of course, it would be a lot more work to develop and maintain at that level. Those two facts are not unrelated.