Air routes where most passengers are connecting?
November 6, 2018 1:57 AM Subscribe
Just out of curiosity, are there any air routes where a high percentage of passengers will be connecting to another flight? Which routes is this common on?
Anywhere into a big hub. Dubai, Hong Kong, Amsterdam-- those are all likely places to have a high percentage of connecting passengers.
posted by frumiousb at 2:12 AM on November 6, 2018 [4 favorites]
posted by frumiousb at 2:12 AM on November 6, 2018 [4 favorites]
There are flights from Albany NY to Newark and/or JFK. Virtually nobody would be on those unless they were connecting. To go to NYC from Albany, Amtrak is much more practical.
posted by jkent at 2:48 AM on November 6, 2018 [1 favorite]
posted by jkent at 2:48 AM on November 6, 2018 [1 favorite]
Anecdote, not data, but I once flew on a completely full flight from Mumbai to Munich. There were two of us (two!) at the Munich baggage claim. Everyone else connected on.
posted by whitewall at 2:49 AM on November 6, 2018 [2 favorites]
posted by whitewall at 2:49 AM on November 6, 2018 [2 favorites]
Any Australian city into an Asian or Middle Eastern hub. It's how we get to Europe.
posted by deadwax at 3:14 AM on November 6, 2018 [2 favorites]
posted by deadwax at 3:14 AM on November 6, 2018 [2 favorites]
Definitely Doha and Dubai, the hubs for Qatar Airways and Emirates, respectively.
posted by EndsOfInvention at 3:36 AM on November 6, 2018 [2 favorites]
posted by EndsOfInvention at 3:36 AM on November 6, 2018 [2 favorites]
I would add Frankfurt, the hub for Lufthansa.
posted by peacheater at 3:59 AM on November 6, 2018 [4 favorites]
posted by peacheater at 3:59 AM on November 6, 2018 [4 favorites]
I can't find numbers, but some of the Essential Air Service routes surely qualify. (It doesn't make a ton of sense to fly from Rutland, VT to Boston if you aren't connecting and there are many EAS routes that are shorter drives than that.)
Similar to Albany - EWR or Albany - JFK, Burlington, VT - LGA is largely people connecting elsewhere. Amtrak takes maybe an hour longer and is much more pleasant, though there are far fewer trains than flights. Because it's a lot cheaper to fly within the US than it is between Canada and US, Burlington also has a good number of people driving from Quebec. Plattsburgh, NY bills itself as "Montreal's US airport" and has flights to Florida on Allegiant and Spirit (for people vacationing in Florida) and to Dulles on United, which will be mainly people connecting.
posted by hoyland at 4:01 AM on November 6, 2018
Similar to Albany - EWR or Albany - JFK, Burlington, VT - LGA is largely people connecting elsewhere. Amtrak takes maybe an hour longer and is much more pleasant, though there are far fewer trains than flights. Because it's a lot cheaper to fly within the US than it is between Canada and US, Burlington also has a good number of people driving from Quebec. Plattsburgh, NY bills itself as "Montreal's US airport" and has flights to Florida on Allegiant and Spirit (for people vacationing in Florida) and to Dulles on United, which will be mainly people connecting.
posted by hoyland at 4:01 AM on November 6, 2018
I read recently that 80% of the passengers who fly out of Charlotte (CLT) are on connecting flights.
posted by Sweetie Darling at 4:11 AM on November 6, 2018 [1 favorite]
posted by Sweetie Darling at 4:11 AM on November 6, 2018 [1 favorite]
A hub is not a route so, e.g.ATL isn’t an answer to the question.
I think your best bet is to look at flights originating in tiny airports, rather than look at what passes through busy hubs.
For example the airport in State College, PA only has (had?) flights to two cities, ever: DC and Detroit. So when you boarded a flight out of town, usually literally every single person on the plane was transferring, because you’re looking at the ratio of those who want to go to the Detroit area to those who want to go anywhere else in the county, and that’s a pretty high percentage on most days.
posted by SaltySalticid at 4:41 AM on November 6, 2018
I think your best bet is to look at flights originating in tiny airports, rather than look at what passes through busy hubs.
For example the airport in State College, PA only has (had?) flights to two cities, ever: DC and Detroit. So when you boarded a flight out of town, usually literally every single person on the plane was transferring, because you’re looking at the ratio of those who want to go to the Detroit area to those who want to go anywhere else in the county, and that’s a pretty high percentage on most days.
posted by SaltySalticid at 4:41 AM on November 6, 2018
Nobody flies from Philadelphia to New York without connecting (you’d literally spend longer dealing with planes than with driving, and if you don’t want to drive you take the train) but the flights exist because Philadelphia and the New York airports are hubs. Also Philly to Baltimore or Washington.
posted by madcaptenor at 4:49 AM on November 6, 2018
posted by madcaptenor at 4:49 AM on November 6, 2018
True, Dubai isn't a correct answer to the question as posed.
But I would look into flights from Northern Indian cities to Dubai and Doha, as well flights from all Indian cities to Frankfurt (and vice-versa). (More people from Southern Indian cities are likely to stay and work in Gulf hubs).
posted by peacheater at 4:50 AM on November 6, 2018
But I would look into flights from Northern Indian cities to Dubai and Doha, as well flights from all Indian cities to Frankfurt (and vice-versa). (More people from Southern Indian cities are likely to stay and work in Gulf hubs).
posted by peacheater at 4:50 AM on November 6, 2018
Also India to Istanbul. I was on an Atlanta to Istanbul flight on Turkish Airlines and many of the passengers were Indian; I assume part of the reason Atlanta has this flight is the large Indian population here using it for connecting. (I was also connecting, but on a route that didn’t make geographic sense, namely to Vienna.)
And anywhere to Reykjavik. Nobody actually lives in Iceland.
posted by madcaptenor at 5:06 AM on November 6, 2018 [1 favorite]
And anywhere to Reykjavik. Nobody actually lives in Iceland.
posted by madcaptenor at 5:06 AM on November 6, 2018 [1 favorite]
Every flight from Lynchburg, VA, goes to Charlotte, NC.
Every flight from Durango, CO, goes to Denver, CO.
Is that what you mean?
posted by Ms Vegetable at 5:51 AM on November 6, 2018
Every flight from Durango, CO, goes to Denver, CO.
Is that what you mean?
posted by Ms Vegetable at 5:51 AM on November 6, 2018
Salt Lake City, a Delta hub.
Miami, lots of flights connecting to the Caribbean, South America, and South Africa.
posted by JoeZydeco at 6:21 AM on November 6, 2018
Miami, lots of flights connecting to the Caribbean, South America, and South Africa.
posted by JoeZydeco at 6:21 AM on November 6, 2018
Rochester International Airport in Minnesota has only three destinations for departing commercial aircraft: Chicago (O'Hare), Atlanta, and Minneapolis. The latter, in particular, is only 88 miles away from RST - close enough that there are companies who operate shuttle buses to drive to Rochester, which is usually quicker. While some of the passengers on the RST-OHD or RST-ATL flights may actually be traveling to Chicago or Atlanta, it's almost unfathomable that anyone departing RST for MSP would actually intend to stop in MSP.
Of note, while the only commercial flights to and from RST are domestic, the airport maintains a Customs post because of the large volume of international private aircraft ("the Saudis", as they're known colloquially) travelling to and from the Mayo Clinic.
posted by kevinbelt at 7:01 AM on November 6, 2018 [1 favorite]
Of note, while the only commercial flights to and from RST are domestic, the airport maintains a Customs post because of the large volume of international private aircraft ("the Saudis", as they're known colloquially) travelling to and from the Mayo Clinic.
posted by kevinbelt at 7:01 AM on November 6, 2018 [1 favorite]
Ah, I guess I misunderstood the question.
Definitely ORD-DEN on United. There's about a dozen flights each day in both directions. Both airports are large United hubs and passengers are connecting between the two to get to other desinations.
posted by JoeZydeco at 9:15 AM on November 6, 2018
Definitely ORD-DEN on United. There's about a dozen flights each day in both directions. Both airports are large United hubs and passengers are connecting between the two to get to other desinations.
posted by JoeZydeco at 9:15 AM on November 6, 2018
Best answer: Inside the US (note: all this is domestic travel only), there's a really good dataset, the DB1B file from the Bureau of Transportation Statistics*. It's a large - 10% - sample of tickets and helpfully includes both information on the "market" and the flights taken; if you fly from, say, Boston to Boise via Denver, you can tell that the true trip is BOS-BOI, but the flights are BOS-DEN-BOI.
I parsed the Q1 2018 file, which has 9 million observations (discarding bulk and group travel), and overall about 53% of passengers on a given flight within the US are connecting. (Note that connecting pax are counted multiple times; if Sam and Diane are both leaving Boston but Sam is going to Boise and Diane to Denver, then he gets counted as a pax on both the BOS-DEN and DEN-BOI flights; she's direct and only counted on the BOS-DEN flight.)
Looking at flights with 10,000 or more pax in the dataset (top 337 flights), the highest connection rates are short flights to major hubs, in the South. #1 is Atlanta-Birmingham (all stats are bidirectional, ie ATL-BHM + BHM-ATL) at 98.3% connecting; of the 15 pairs with the highest connecting rates (90%+), 5 are to Atlanta (Birmingham, Charleston, Greenville SC, Pensacola, Savannah), 6 involve Charlotte (Charleston, Wilmington NC, Jacksonville, Norfolk, Richmond, Raleigh-Durham) and the other 4 involve Texas hubs (Austin to Dallas and Houston, OKC to Dallas, San Antonio to Houston).
The same pattern is true outside of the south even though the top 15 are all down there; LAX-San Diego is 89.3% connecting, Anchorage-SeaTac is 83.9% connecting, Indianapolis-O'Hare is 83% connecting. Looking at pairs with 5K-10K pax in the dataset, you see similar trends; Colorado Springs-Denver is 99.5% connecting, Grand Rapids-Detroit and Milwaukee-O'Hare are both 98.6%; Fargo and Sioux Falls to MSP are both in the 93% range. (And a ton more cities to Atlanta or Charlotte).
The overall highest volume of connecting passengers within the US is on the Atlanta-Orlando flight, followed by Atlanta-Tampa and Atlanta-Ft Lauderdale flights; these are in the 70s percent wise, but are also high volume flights. There's a ton of demand to go to Florida (NB: this is the Q1 file, so it has particularly high seasonal demand), and a ton of people who can connect via Atlanta but might not have a direct flight to Florida.
This is borne out by the flights with the lowest share of connections; they're primarily from the Northeast to Florida, particularly where one of the ends is a second-tier airport; Cleveland to Fort Myers is the lowest flight with 5K+ pax at 1.5% connecting; Boston to Fort Myers is in second (it's #1 for flights with 10K+ pax) at 2.3% connecting; the rest of the top 15 includes Boston to Ft Lauderdale, Palm Beach, Orlando and Tampa; on the other end, Orlando flights from Bradley (northern CT), Macarthur (aka Islip, Long Island), Providence and Milwaukee.
Flights between big hubs tend to have a lower rate of connecting passengers (I'll define big hubs as the top 20 airports, which includes Salt Lake City or larger); around 45% instead of the 53% in the whole dataset -- not to pick on JoeZydeco, but the ORD-DEN flight is only 55% connecting because I suspect there's plenty of people who actually want to be in Chicago or Denver, and also because there's plenty of options for connections; if you're going between mid-size cities (eg Boise or Albuquerque to Madison or Columbus) you can probably do a single-hop connection via ORD or DEN without having to go through both. (If you're going between tiny cities like Helena to Lansing, you connect twice but it's not enough people to count.) The highest rates of connecting pax between big hubs are FLL-MCO (83.1%), ATL-SLC (79.2%), DTW-SLC (76.3%), DFW-IAH and DTW-SEA, (both 73.4%).
In fact, it looks like Delta takes their hubs more seriously than United or American; of the top 20 highest connecting rates for pairs involving these big hubs, 8 are between Delta hubs (ATL, DTW, SLC, MSP, SEA) and another 5 have one end at a Delta hub.
To sum up, in the US, the highest rates of connecting flights are shorter distance flights between a small airport to a major hub. The short distance means a high share of the people who want to go direct can drive; the small airport means it has few alternatives for direct flights; the major hub means it has lots of opportunities for connections.
Extrapolating to intercontinental flights, I suspect the highest connection rates are between smaller cities and major hubs, particularly hubs that have a disproportionate size -- Doha > Singapore > Tokyo; Reykjavik > Munich > London.
* One of many government agencies under siege, remember to vote if you can!
posted by Homeboy Trouble at 10:37 AM on November 6, 2018 [34 favorites]
I parsed the Q1 2018 file, which has 9 million observations (discarding bulk and group travel), and overall about 53% of passengers on a given flight within the US are connecting. (Note that connecting pax are counted multiple times; if Sam and Diane are both leaving Boston but Sam is going to Boise and Diane to Denver, then he gets counted as a pax on both the BOS-DEN and DEN-BOI flights; she's direct and only counted on the BOS-DEN flight.)
Looking at flights with 10,000 or more pax in the dataset (top 337 flights), the highest connection rates are short flights to major hubs, in the South. #1 is Atlanta-Birmingham (all stats are bidirectional, ie ATL-BHM + BHM-ATL) at 98.3% connecting; of the 15 pairs with the highest connecting rates (90%+), 5 are to Atlanta (Birmingham, Charleston, Greenville SC, Pensacola, Savannah), 6 involve Charlotte (Charleston, Wilmington NC, Jacksonville, Norfolk, Richmond, Raleigh-Durham) and the other 4 involve Texas hubs (Austin to Dallas and Houston, OKC to Dallas, San Antonio to Houston).
The same pattern is true outside of the south even though the top 15 are all down there; LAX-San Diego is 89.3% connecting, Anchorage-SeaTac is 83.9% connecting, Indianapolis-O'Hare is 83% connecting. Looking at pairs with 5K-10K pax in the dataset, you see similar trends; Colorado Springs-Denver is 99.5% connecting, Grand Rapids-Detroit and Milwaukee-O'Hare are both 98.6%; Fargo and Sioux Falls to MSP are both in the 93% range. (And a ton more cities to Atlanta or Charlotte).
The overall highest volume of connecting passengers within the US is on the Atlanta-Orlando flight, followed by Atlanta-Tampa and Atlanta-Ft Lauderdale flights; these are in the 70s percent wise, but are also high volume flights. There's a ton of demand to go to Florida (NB: this is the Q1 file, so it has particularly high seasonal demand), and a ton of people who can connect via Atlanta but might not have a direct flight to Florida.
This is borne out by the flights with the lowest share of connections; they're primarily from the Northeast to Florida, particularly where one of the ends is a second-tier airport; Cleveland to Fort Myers is the lowest flight with 5K+ pax at 1.5% connecting; Boston to Fort Myers is in second (it's #1 for flights with 10K+ pax) at 2.3% connecting; the rest of the top 15 includes Boston to Ft Lauderdale, Palm Beach, Orlando and Tampa; on the other end, Orlando flights from Bradley (northern CT), Macarthur (aka Islip, Long Island), Providence and Milwaukee.
Flights between big hubs tend to have a lower rate of connecting passengers (I'll define big hubs as the top 20 airports, which includes Salt Lake City or larger); around 45% instead of the 53% in the whole dataset -- not to pick on JoeZydeco, but the ORD-DEN flight is only 55% connecting because I suspect there's plenty of people who actually want to be in Chicago or Denver, and also because there's plenty of options for connections; if you're going between mid-size cities (eg Boise or Albuquerque to Madison or Columbus) you can probably do a single-hop connection via ORD or DEN without having to go through both. (If you're going between tiny cities like Helena to Lansing, you connect twice but it's not enough people to count.) The highest rates of connecting pax between big hubs are FLL-MCO (83.1%), ATL-SLC (79.2%), DTW-SLC (76.3%), DFW-IAH and DTW-SEA, (both 73.4%).
In fact, it looks like Delta takes their hubs more seriously than United or American; of the top 20 highest connecting rates for pairs involving these big hubs, 8 are between Delta hubs (ATL, DTW, SLC, MSP, SEA) and another 5 have one end at a Delta hub.
To sum up, in the US, the highest rates of connecting flights are shorter distance flights between a small airport to a major hub. The short distance means a high share of the people who want to go direct can drive; the small airport means it has few alternatives for direct flights; the major hub means it has lots of opportunities for connections.
Extrapolating to intercontinental flights, I suspect the highest connection rates are between smaller cities and major hubs, particularly hubs that have a disproportionate size -- Doha > Singapore > Tokyo; Reykjavik > Munich > London.
* One of many government agencies under siege, remember to vote if you can!
posted by Homeboy Trouble at 10:37 AM on November 6, 2018 [34 favorites]
That's insanely impressive. Wow.
posted by JoeZydeco at 10:47 AM on November 6, 2018 [1 favorite]
posted by JoeZydeco at 10:47 AM on November 6, 2018 [1 favorite]
Thanks, Homeboy Trouble! I was just starting to think "this data's got to be out there somewhere..." and someone actually did the work.
When you say flights with more than 10,000 pax in the dataset - are we talking flights with about 40,000 pax per year (since it's a dataset for a quarter) or about 400,000 (since it's also a 10% sample)?
posted by madcaptenor at 12:01 PM on November 6, 2018
When you say flights with more than 10,000 pax in the dataset - are we talking flights with about 40,000 pax per year (since it's a dataset for a quarter) or about 400,000 (since it's also a 10% sample)?
posted by madcaptenor at 12:01 PM on November 6, 2018
When you say flights with more than 10,000 pax in the dataset - are we talking flights with about 40,000 pax per year (since it's a dataset for a quarter) or about 400,000 (since it's also a 10% sample)?
400,000 -- subject to the caveat that Jan-Mar travel has patterns that are different than the rest of the year.
While I'm at it, here's a pastebin link that has all the airport pairs with 2500 pax (ie ~100,000 / year bidirectional) split into direct, connecting and percentage.
posted by Homeboy Trouble at 12:39 PM on November 6, 2018
400,000 -- subject to the caveat that Jan-Mar travel has patterns that are different than the rest of the year.
While I'm at it, here's a pastebin link that has all the airport pairs with 2500 pax (ie ~100,000 / year bidirectional) split into direct, connecting and percentage.
posted by Homeboy Trouble at 12:39 PM on November 6, 2018
« Older Software you paid for or consider paying for? | Hull painting scheme of SSBN-643 USS George... Newer »
This thread is closed to new comments.
posted by twirlypen at 2:06 AM on November 6, 2018 [2 favorites]