I've decided to make a separate thread to host my blathering about Brightline's ridership statistics (such as I can generate them) and not overwhelm the main thread. Some notes to start:
Each Brightline train, at present, contains the followjng
Car 1: 48 Select seats, 1 wheelchair slot
Car 2: 32 Smart seats, 32 SmartPlus seats, 2 wheelchair slots
Car 3: 64 Smart seats, 2 wheelchair slots
Car 4: 56 Smart seats, 2 wheelchair slots
-152 Smart seats
-32 SmartPlus seats
-48 Select seats
-7 wheelchair slots
--Overall total: 239 seats and slots (232 for purposes of my analysis)
Unless or until I can get someone to help me mount a program on a test harness to run checks, I have to basically use a clicker on the Brightline website and go through each day's trains to get a sense of ridership. With about 14-16 trains per day, this represents 84-96 separate checks I have to make for a full day of trains. A single run-through therefore takes about an hour, especially since on a packed train I may need to count Smart 2-3 times to be comfortable with the count. This means that I am not going to turn around and bother with the wheelchair spaces (I am ignoring them for my purposes). It also means that, in the vast majority of cases, I am only going to check Palm Beach-Miami or vice-versa, since checking to/from Fort Lauderdale would triple the workload. In some cases I may try to make a note of extra sales on a train between the Miami/Palm Beach departure and the sales cutoff for Fort Lauderdale. For example, on one train yesterday morning (the southbound 0700) I noted that an additional twelve seats were sold FLL-MIA after the train left West Palm.
Some things I have observed, on what is thus far a sample size of one day imperfectly observed:
-Somewhere around 30% of tickets yesterday were "walk-up" sales, being sold in the 60-90 minutes before a train departs. This tends to be a little bit lower for peak trains (e.g. peak hours in the main direction of flow) but higher for off-peak trains. Day-of sales tend to be higher for trains later in the day as well (no shock there), in at least one case exceeding 50%.
-Load factors yesterday were reasonably healthy at rush hour. Outside of rush hour...not so much. Rush hour trains seem to have load factors in the 45-70% range in the peak direction, just under 20% in the reverse-peak direction, and often 5-10% in the off hours. I observed some even lower numbers, but per the walkup ticket situation they likely came in a bit higher than I saw. Overall load factor was probably about 13-15% (I technically observed 11.4% northbound and 13.5% southbound , but as noted above I also probably missed out on about 100 ticket sales from my count within the methodology I was following.
--I may well have missed out on another 50-100 due to my methodology not being able to fully pick up either last-minute sales out of Fort Lauderdale or any cases where a seat is sold both Miami-Fort Lauderdale and Fort Lauderdale-Palm Beach (or vice-versa). The former is likely to show up in a lot of cases; the latter is probably only a notable concern at rush hour (particularly heading south, since I presume that Fort Lauderdale is a net-pickup station SB and a net-discharge station NB).
-There are some cases where large blocks of seats sell out together, quite often in Car 4. I do not know if that is because of group sales, automatic sales in the system, or if there are "phantom sales" showing up.
 The southbound 0530 train did not run yesterday.
Edit: Explanation of charts:
-Each line is a departure from one of the origin stations (West Palm for southbound, Miami for northbound).
-I've used "F" to indicate Select, "W" to indicate SmartPlus, and "Y" to indicate Smart. This is based on airline coding practices (since being able to use a single letter makes the chart cleaner).
As further explanation, I will generally confine the data on a given day to a single post, even if I ultimately post 2-3 sets of data in here. For days I cover (which won't be all days), I'll try to have a "final set" (which will necessarily be incomplete) and at least one full "snapshot set" (usually taken between 2200 and 0400 the night before). I may occasionally keep another set from a day or two before as well, particularly if I think traffic is going to be particularly heavy on a given day.
Edited by Anderson, 20 September 2018 - 04:47 AM.