Animated flight maps


#1

Make the flight maps like weather rader where you can see the last 5-10 updates. Basically make an animated GIF.

This way we can watch the little planes move :smiley:


#2

We make animations for special events (usually aircraft circling for hours due to gear problems), but unfortunately it’s just too computationally expensive to create one for every flight.


#3

Sorry, I should have been more specific.

I meant for the Airport overviews. It would be great to see ORD for the last 1/2 hour. As far as expensive, I don’t see how this could be. All that is need ed would be to modify the previously saved animated gif in a First In First Out fasion. I can see this as a only being more expensive on the order of N, and as we (computer scientists) know, N is good, cheap and fast.

food for thought.


#4

I don’t think mduell meant $$$ expensive- he meant the processing needed to aggergate the flights for the 10 minute loop, pull all the positions, create the maps, and send them as a loop requires a tremendous amount of computational power(remember that many hundreds of people would be looking and various different movies at the same time).

Even if you developed a little java(or such) applet that would let the client draw the final maps, ect, you still would be pushing much more data out of FA. Best case might be an applet that collect the positions as they are updated in real time, then allow the user to “replay” the previous 10 minutes.

Would be cool.


#5

Also, expensive can be in terms of labor/time costs. These computer geeks (nothin but love for ya!!) are NOT cheap!


#6

I didn’t mean expensive ‘$$$$’-wise.

Workload in processing is usually discussed using the term O(N), or order of N. In terms of efficiency, O(1) is the best, O(N) is great, O(n^2) is exponentially worse. Again this represent the cost associated with processing needed. Now, back to my example.

In my example, the app server would collect the data for an airport every 90 seconds, or what ever it is right now. When this happens, the app server would take that info for the current plots and create a new frame for the animated gif, paste it into the gif and then eliminate the oldest gif, in this case the one that is ten minutes old. This is how animated weather maps on sites like weather.com, chicagotribune.com, etc work.

When the client requests the latest a greatest from the server, it grabs the latest animated gif and ZERO processes is done on the client, just like it is now.

Now back to the efficiency: This is O(N) because the server only has to do this for the number of airports in the world. (It is actually O(1), but I am assuming that the number of airports could change down the road). But the bottom line is that the processing needed is LINEARLY affected based on the number of airports, not EXPONENTIALLY. In the computer world, we say this is “good, fast, and CHEAP”


#7

You are assuming the map servers create maps for all airports, save them, then issue the pre made display when requested. I do not believe* this is the case- the map servers create the maps on the fly when a request comes in. This means we only make maps that are actually needed**, not a map for every airport in the world***

Your proposal would require the maps server to constanly make and store maps for every airport (10xN) , which would certaintly require more power.

  • See THIS
    ** I don’t know if maps are cached. I expect at busy airport (say ORD) get many more requests than say Moraine Airpark. If I (mapserver) just made a ORD map .05 secs ago, do I make another? (Probably).

**Or FAA controlled airspace

  • Edited to add I still think it would be a really COOL feature- I would love to watch a busy airport reconfigure runways after a wind change, or watch the action after a TFR for a VIP airport visit is lifted.

.


#8

JReeves -

A couple of notes:

O(10xN) == O(N) in terms of processes because it is still linear. And like I said before, it is really O(1) (which equals O(10) for the same reason as above except instead of linear, it is now static.)

Also, I don’t believe that the site is processing the map on each request. The reason I believe this is because if I refresh (F5) the page, (even full refreshes (ctrl+F5)), the map doesn’t change most of the time. It only seems to change every 60-90 seconds, or so. Now having said that, and I mean NO disrespect to the developers at FA, but if they are actually processing the map EVERY request, which I doubt, that is extremely inefficient and in the future, the site will strain under the request load. (assuming the site gets mega-popular, which would be awesome)

Again, I think the best comparision is animated national weather maps. There is not one radar watching the entire country, there are many, and ever so often a server stiches together all the radars and creates a map. This is then applied to a animated gif in a FIFO fashion thereby creating our lovable animated maps. The stitching and applying to the gif happens on a fixed schedule.


#9

You likely will not see a difference in maps when refreshing @ 60s intervals, as this is best case resolution of data points comming from FAA, plus a little process time to put the in the active database for the map servers to access.

Based on some fast estimating, I figure there are about 5,000 airports in the 50 US states (see Airnav Airport Browse by State - quick conservative* look is about 100 airport per state * 50 states for about 5,000. (If anybody has a better number, I would be interested in knowing. I assume that the FA database contains more than this number.

As the number of users actively refreshing (N1) approaches number of airports (N2), the strategy of calcing all airports and serving the results does become more processor friendly. However:

  1. The maps would become more aged- you would hold the map image in storage to serve requests for a period of time (1min or greater I would assume) this would add on to the delays already imposed by FAA (5-6min)

  2. My understanding is the FA plans to allow more flexible map centering as part of an upgraded mapping project (center on Navaids, fixes, or (hopefully) any user given points. In this case possible values for N2 DO go up exponintially.

A more sane** request would be to pick a few "exciting airports that very likely do get queried every minute or so, save the image from them and allow a moving map of those airports.

    1. I know that this data is available Here on FA, but the airnav site shows the airport count by state on the index page, thus I linked it.
      2 I hate being conservative.

** I am not questioning your sanity! I mean sane in the computer processing sense! - by the way: Welcome to FA- hope you have a long, pleasant flight!


#10

OOOOOOOOWWWWWWWWWWWW!!!

Will you two stop already? You are making my brain hurt!!!


#11

Your Brain hurts? It’ll hafta come out. I know. I’m a brain specialist. Ooh! We forgot the anaesthetic!


#12

jreeves,

The number of computations needed to serve up a dynamic map is soley a function of the number of users*, not the constant** number of fixes/airports/navaids/lat+lon for which data needs to be polled. The only difference is the lat/lon of the “center”, as well as how far way (+/- n degrees lat/lon) aircraft to be included are.

*another browser window is another user

**The switchover point to the next AIRAC data is simply a change to the constant

I agree there should be a default time lapse for the larger airports, and that for smaller ones users should initiate the request to animate the map (start the FIFO queue)


#13

How about a compromise here?

Could FA push out each airport map update into a file on the user’s computer desk top (just like an RSS feed) and the user’s computer would then do the animation via QuickTime or some other ‘video’-generating software?

No animation computation for FA. Just a picture pushed to the user as long as his browser remained connected to FA. It would be the same picture that FA updates for its map display anyway.


#14

the idea for having it like the weather radar on tv, that would be cool to see that on say the whole country and just busy airports. But you should keep that as an option to show when you choose an aiport to view.


#15

Can’t a person simply run an auto program such as “automize”, grab the images, rename them with a time stamp, then send to a server for a looping software to configure the images for a movie? I’ve seen this done on youtube with flightaware before, though I’m not clear if that person was manually saving the images, or auto downloading them. Anyone try this before?

Jon


#16

Can’t a person simply run an auto program such as “automize”, grab the images, rename them with a time stamp, then send to a server for a looping software to configure the images for a movie? I’ve seen this done on youtube with flightaware before, though I’m not clear if that person was manually saving the images, or auto downloading them. Anyone try this before?

Jon


#17

I believe there is a moving map for the Atlanta area. However, if you know what you want and how to make it happen, why not create this yourself?


#18

I never thought to look up flightaware on youtube! Here’s Daniel sounding all professional-like on the phone. :slight_smile:
https://www.youtube.com/watch?v=kd_Vxub6bqU)


#19

I know what I want to do, but don’t know how to do it. It sounds good on paper. :slight_smile:


#20

Not to market the competition, but this JAVA animated map is not too shabby: www4.passur.com/sgj.html

I’m sure all these companies get their data from the same source…it’s just a matter of getting the resources to do something like this on a nationwide scale (and still keep the service free!) :wink: