I use FlightAware
all the time, but didn't know they pull the data directly from the planes.
I'm guessing FlightAware is one half of the "duopoly" spoken about in the submission.
Edit:
Looks like the cost per 1k queries for flight arrival times is $7.90 USD, which is a Class 2 query @ $0.0079 USD per query at the highest tier [0], but it goes down with more queries. You actually also probably want to pull their "AirlineFlightSchedules" API data to pull initial schedules for the day, which is a Class 1 query @ $0.0120 [0].
How many flights globally, per day? ~100,000 [1]. Once you fall down into their tiered pricing, and use their "15 results per query" rule, it amounts to ~$80. Once there is flight information for each flight for the day, there will likely be periodic update times. There's a firehose feed, but they don't specify pricing, so I'm not sure what the cost is to provide live updates.
If we say 5 updates per flight (total guess), that becomes ~500,000 queries daily of Class 2 data. I'm confident there's a way to batch these, so this comes down to ~33,334 requests, which puts us at ~$181.34 in additional charges per day.
So our daily costs are now about ~$261.34, or ~$7,840.20 per month, assuming they're not pulling data on-demand. This changes somewhat significantly if an app (correctly IMO) only pulls data for flights it's asked about. You get to cut out the $80/day entirely, and only some portion of the $181.34 is actually paid each day. If you consider the ~28,500 domestic flights in the US [2], and if app users only search for half of those, costs drop to ~$162 per day, or $4,860 monthly (plus the unknown cost of the firehose).
So, because of the cost of the data alone, your app will have to pull in a bunch of users. Wait, how many? Let's figure that out, based on an estimation strategy I pulled off of Quora (so it might be wrong) [3]:
Each user checks their flight status 3 times, which takes them ~1 minute overall to read, so we get 2 impressions [3]. I don't really know what these other two things are, but I guess we'll next assume 100% fill rate and eCPM of $3. But that's just per flight. 694 million passengers took flights in the US this year [4], and while I know we're getting into some pretty speculative territory, that's at least 2 flights per person per year, not to mention all the other people who also check on flight statuses of their friends/family. I feel comfortable saying that each person would check this app at least 15 times a year (you'd use it if you had it), or 0.04 times a day, which would give 0.08 impressions per day.
Our calculation is now:
(0.08n / 1000) * $3 * 1.0 = $0.0002n
$0.0002n = $162
n = 810,000
To cover costs for the data alone, you'd need 810,000 users. Even if each person used this app twice as often, you'd still need 405,000 users. That's completely untenable. I'm actually wondering if my math and assumptions are correct...
[0] http://flightaware.com/commercial/flightxml/pricing_class.rvt
[1] http://www.garfors.com/2014/06/100000-flights-day.html
[2] http://sos.noaa.gov/Datasets/dataset.php?id=44
[3] https://www.quora.com/How-much-ad-revenue-can-be-expected-per-100-000-downloaded-iPhone-iPad-apps
[4] http://www.transtats.bts.gov/