Fascinating as usual. The danger of the flight path convergence is so obvious. No matter what the cause of the Regan incident, the convergence provided the opportunity for any accident do close to the airport. The Philadelphia accident has to be mechanical issues. What is the detailed process for inspections? Was the pilot blameless?
Hey Chip - thanks for reading! I completely agree. In the Philadelphia incident- I haven't heard much conclusive - but one theory is that cargo came lose on take off and shifted to the back of the plane, upsetting the attitude, and eventually stalling the jet. We'll have to see what comes out in the investigation though. Hope all's well!
Excellent work! Your visualizations and descriptions make this problem and a path forward (no pun intended) perfectly clear. I hope news sources and ultimately, the folks in charge of re-evaluating flight paths that clearly increase the likelihood of incidents, will find and use the resource you've created here!
Thank you so much! My hope is that there are professionals out there already thinking deeply about these problems, and employing systems thinking like this. This example is how I might approach the problem, but I really hope someone else is already on it. I really appreciate you reading...as always!
Nice visualization. I understand it's based on simulated data converging at the runway, but even in this case, what proportion of the paths intersect (within some spatial tolerance)?
What would it take to get actual or historical transponder data? It would be interesting to compare this generated data to actual transponder data.
I agree with your note that looking at the density of "intersections" among all transponder data could highlight routes with especially problematic proximity/density in space-time. Does FAA currently do anything similar for route planning or updates?
All good questions - I did not do an analysis of the intersections of the paths ... since most of this would have been a result of my "randomization" params. I did want to sort of normalize this data and create "slices" that would represent probabilities that an area of space would have an aircraft in it over a given period of time... and then use these areas to look for problematic intersections. However, all needs real data.
To get real data, it's not hard, but you have to have a paid business tier on flightradar or other providers, so it's not cheap. But, they do have lots of it.
I would hope the FAA/NTSB has some sort of analysis like this going on. They would have even more rich data than the transponder data publicly available. Also, a real data scientist could probably come up with a better method looking purely at the numbers - this was more an exercise for me in using visualization tools (in javascript) that I was familiar with.
Fascinating as usual. The danger of the flight path convergence is so obvious. No matter what the cause of the Regan incident, the convergence provided the opportunity for any accident do close to the airport. The Philadelphia accident has to be mechanical issues. What is the detailed process for inspections? Was the pilot blameless?
Thanks
Chip Newhart
Hey Chip - thanks for reading! I completely agree. In the Philadelphia incident- I haven't heard much conclusive - but one theory is that cargo came lose on take off and shifted to the back of the plane, upsetting the attitude, and eventually stalling the jet. We'll have to see what comes out in the investigation though. Hope all's well!
Excellent work! Your visualizations and descriptions make this problem and a path forward (no pun intended) perfectly clear. I hope news sources and ultimately, the folks in charge of re-evaluating flight paths that clearly increase the likelihood of incidents, will find and use the resource you've created here!
Thank you so much! My hope is that there are professionals out there already thinking deeply about these problems, and employing systems thinking like this. This example is how I might approach the problem, but I really hope someone else is already on it. I really appreciate you reading...as always!
Nice visualization. I understand it's based on simulated data converging at the runway, but even in this case, what proportion of the paths intersect (within some spatial tolerance)?
What would it take to get actual or historical transponder data? It would be interesting to compare this generated data to actual transponder data.
I agree with your note that looking at the density of "intersections" among all transponder data could highlight routes with especially problematic proximity/density in space-time. Does FAA currently do anything similar for route planning or updates?
All good questions - I did not do an analysis of the intersections of the paths ... since most of this would have been a result of my "randomization" params. I did want to sort of normalize this data and create "slices" that would represent probabilities that an area of space would have an aircraft in it over a given period of time... and then use these areas to look for problematic intersections. However, all needs real data.
To get real data, it's not hard, but you have to have a paid business tier on flightradar or other providers, so it's not cheap. But, they do have lots of it.
I would hope the FAA/NTSB has some sort of analysis like this going on. They would have even more rich data than the transponder data publicly available. Also, a real data scientist could probably come up with a better method looking purely at the numbers - this was more an exercise for me in using visualization tools (in javascript) that I was familiar with.
Thanks for reading! Appreciate all the thoughts.