dfwjim1 wrote:Since the airlines are sending in the data is it possible that they could be sending in false numbers to make the LFs look better than they really are?
It is entirely possible, however having done some of this work myself in relation to BOS, I've spoken to the in-house person who does the work for Massport and to be honest they are not that far off. However one HUGE caveat that I was told with the T-100 data is that it only reports revenue passengers, not ALL passengers, so these numbers will be a little on the lower side.
In addition the reporting can a bit hit/miss. I found gaps in the data for DY back in 2014, EI had a bad couple of months around June last year, even BA missed a month once.
In terms of the data files, it all depends on which database you actually use, as they are slightly different. Sometimes the International only database does not contain all the international pieces (i've found that out before) and the main database includes diversions, some of which are easy to figure out, some of them are not, so trying to figure out route metrics is not great. Lastly 2 airlines have major issues in reporting, they are QR and EI. QR only reports everything as a 773, when it's been a 350 to BOS pretty much since inception apart from a couple of months, also their seat counts don't always match as a result, forcing incorrect %'s if you don't review. EI has a similar problem, they list everything as a 332, even though we know they fly 757's and 333's at the least and the odd leased in 767 as well. Their seat counts also don't always tie very well, meaning they are shorting it by having 1 line as the 332, but reporting a mix of 757's and 333's in there too. I've tried analyzing using thelingussource.com (which is a great resource site) and i've still not really been able to figure out exactly what they have reported.
Upshot, the T-100's are what they are, they are a snapshot, are they the most accurate numbers, no absolutely not and all quoting of numbers from them should be taken a grain of salt and should be considered directional at best. Some airlines are very good and accurate I feel, B6 for example is almost always spot on the seat counts due to their limited aircraft types, but others are all over the place. My feeling is that they are good as somewhat of a guide, but they cannot be used to extrapolate true performance. They are the only statistics this comprehensive that are available for free and that's why a lot of folks use them (myself included) but it comes at a cost and let's not forget they speak nothing to yields or profitability either and nor should they. The only sense you can get from that is the limited data that comes from a different data table that analyzes a portion of ticket prices between city pairs. But even in those cases it's tough to narrow down, because they are grouped in a lot of cases by market, not by airport.
Hey, I am not knocking the OP, because I love this stuff and have spent a lot of time working on my own versions relating to BOS and its arriving airlines, but like many others have said in the past you have to be careful drawing conclusions based on these stats, there are just too many wildcards and variables involved.
