So, a friend of mine was curious about this, and now I am as well.

When United.com (for instance) reports that a flight is delayed by 15 mins, where is that data pulled from? Is it pulled directly from internal United.com computer systems - or something else?

Secondly, when you use a third party website - where is that data pulled from? For instance, the flight in question was running late, however they were circuling 1 mile over ORD - and showing an arrival time of 1 minute later, despite the fact the altitude was near 10,000+ feet. So, clearly it seems it's doing simple math to find out the arrival time and not looking at what the airline is posting. That's because the airline (United) was posting a late arrival of about 15 mins, while the 3rd party site was posting an on time arrival.

Further, when 3rd party site shows altitude/speed/etc... is that coming from the FAA (delayed by 10 mins), or the airline? I would assume the FAA since you can get it for virtually every flight, but then that brings up the question of filing a flight plan. If a flight is late by 2 hours, then why does the 3rd party site not show the delayed arrival time, meanwhile United does?

So, I guess I'm curious about the following:

1) Where United.com (or similiar) pulls thier information from
2) Where 3rd party sites pull thier information from
3) Where does alt/speed/etc information come from
4) How does the flight plan filed with the FAA play into the 3rd party sites arrival/departure times?

Thanks!