frequentfryer From United States of America, joined Nov 2012, 30 posts, RR: 0 Posted (4 months 4 weeks 1 day 4 hours ago) and read 1101 times:
I'm dealing with a bit of data overload regarding open source O&D data for the domestic US. I hope this thread will assist the next person learning how to pull O&D data as well.
I want to see the daily average number of passengers traveling directly from origin(a) to all direct destinations by quarter with average fares.
On one end of the spectrum there is a tool on the Airline Empires site called "paxdata" that has a very simple UI and maps the data for you, but customized queries do not seem to be functioning and It appears that you can't look at data over time. Also did not see documentation on how timely the data is.
Transtats is highly sited here. I am attempting to decode the process to get the information I am looking for. I'm sure it looks intuitive to someone already familiar with the too. I'm not sure if I need to build out my own access DB and insert their tables or if there is a way to select a city and timeline and export a spreadsheet with relevant O&D data.
tdscanuck From Canada, joined Jan 2006, 12709 posts, RR: 81 Reply 2, posted (4 months 3 weeks 5 days 2 hours ago) and read 977 times:
Quoting jetmatt777 (Reply 1): Table 6 has every U.S. Domestic city pair with more than 10 pax per day avg. This is total traffic, divide the number by 2 to get an each way estimate (but not precise)
Just be very careful that that the DOT data is for individual legs, not O-D markets. There is no rigorous way to figure out how much of, say, Allentown-DFW traffic is really in the Allentown-DFW O-D market and how much is in the Allentown-Los Angeles market (with a connection in DFW).
Sometimes you can identify routes where this is an issue by major fare discrepancies...for example, the average fare on Allentown-Detroit is $346...there is zero chance that much of that traffic is really just going to Detroit (it's only a one hour flight and it's not that competitive so the fare is too high for just Detroit)...most of those people are connecting through Detroit to somewhere else.
jetmatt777 From United States of America, joined Jun 2005, 2660 posts, RR: 37 Reply 3, posted (4 months 3 weeks 4 days 21 hours ago) and read 946 times:
Right, which is why I said city pair and not non-stop market.
The data would be too large and complex if O/D was specific to each route, flight, and carrier. The city pair, passenger traffic, and market share coupled with some other data (schedules, etc.) should give a fair representation of that route, but by no means scientific.