At the moment AI live traffic is a bit busted, but I understand other things have taken priority. Yet, I appreciate the fact that Microsoft has committed multiple times to having every aircraft in the sky simulated by AI down the line. So this is a wish for the future more than for “right now” (of course, the earlier the better).
I’ve spent over a year observing AI by now since it has been a major area of interest for me since FS2004. There’s something about bringing airports to life that is both relaxing and exciting.
Preamble aside, I’ve come to realize that Microsoft’s promise simply can’t be achieved with the current partner (Flightaware) alone.
We all know that Asobo has been contending with flights that often don’t appear because of lack of data like altitude and more. They’ve made attempts to interpolate it, but it’s still a problem area.
Unfortunately, the problem with Flightaware is deeper.
They have entire fleets of many airlines completely wrong, listing the wrong variants of an aircraft. For instance, ANA’s DH8Ds are wrongly identified in as DH8Bs
Another issue is that many fleets don’t have the specific variants. For example, most flights by Solaseed Air are marked as being flown by generic “737” while everyone able to use Google knows that the airline only has B738s.
Other fleets that have a sizable number of NEO A320s and A321s simply ignore their existence, marking them all as CEOs.
This is not nitpicking. Issues like these are absolutely pervasive in many areas of the world, across many airlines, and would absolutely wreak havoc on any attempt to correctly portray air traffic once Microsoft implements correct models and liveries.
Sticking to the example above, the aircraft labeled as “737” (or other generic variants like 787 and more) would cause the simulator not to know what model to spawn. Wrong variants would confuse it by trying to associate the wrong model with a texture that does not exist.
The worst issue is that from my observation Flightaware does not act on reports of bad data or does so in a direly insufficient way. I’ve reported many, many issues like the above over several months, and none, none has been fixed.
Yet, there’s more that would cause problems in the correct portrayal of worldwide traffic, and it’s the fact that Flightaware does not provide (at least to free account) tail registration numbers.
This means that simulating all those spiffy special liveries correctly would be impossible. An additional problem is that the simulator would have no way to distinguish between Pax and Freight versions of the same aircraft when they operate under the same airline and ICAO/IATA code, which is fairly common with several airlines.
My conclusions after months of observation is that Flightaware alone is simply not sufficient to get a precise picture of worldwide traffic, and as such, it’s not suitable for the needs of Live Traffic of Microsoft Flight Simulator. It’s not suitable now, and it’ll be even less suitable when Microsoft tries to keep its promise to implement its own models and liveries.
I’ve also observed Flightaware’s main competitor, FlightRadar 24, and I have seen that (while not 100% perfect), it’s much, much more complete and precise in the data it provides.
99% of fleets are correctly labeled with the correct types, and registration numbers are provided to all users, permitting a much more granular and specific implementation of air traffic.
On top of this, one of the major weaknesses of the current live traffic in Microsoft Flight Simulator is that it does not even try to simulate the aircraft on the ground at each airport. They spawn completely at random, with the result that you’ll find Easyjets in Hong Kong and so forth. This is likely due to the fact that Flightaware provides no such data at all.
FlightRadar 24 provides comprehensive and updated lists of aircraft on the ground at each airport, which would be invaluable to Microsoft Flight Simulator. While it’s true that the detailed list is available only to Silver members, I would be surprised if Microsoft couldn’t manage to get FR24 to provide it to the simulator.
This means that a partnership with FlightRadar24 would bring much more detailed, correct, complete, and simply higher quality data for aircraft on the air, and potentially data on the ground that is now completely missing.
This is not to say that Microsoft should completely do away with the partnership with Flightaware (albeit if they had to choose between the two for any reason, I’d advise moving to FR24).
The best option would be using data from BOTH services and interporlate it, so that when one service fails, the other can integrate what is missing.
This would result into basically every flight correctly showcased and simulated in Microsoft Flight Simulator, both in the air and on the ground, which is exactly what AI live traffic should do.
Not only it’d benefit the base simulator, but it’d also provide third-party AI developers with a much more solid base to work with.
Thanks for your attention.