Re: Observations from a Demand Forecaster
Author: BOB2
Date: 09-08-2017 - 12:30
1800 and change for the opening day of regular service isn't too bad. The projection of 3,000 seems achievable. But, be careful when using "forecasts" to assess performance. I was chided on the opening of the first leg of the Gold line that it was coming in "under" forecast, as well. However, the forecast was run using the "assumption" of 6 minute headways, when, "in fact" the were operating at 10 minute headways.....
Most transit models used for this kind of forecasting of ridership are very sensitive to the "disutility" of wait time, which is weighted by the "consumer" usually around 3 times the perceived time "in-vehicle" (consumers view sitting around waiting for the bustrain/ with the thumbs up their a$$, as 3 times worse, than the time they spend traveling on the bus/train), which was the about the differential was in the "forecast" versus "actual" counts. Worse, MTA's estimates and counts of "actuals" were so poorly done on the Gold Line, that a subsequent "on board" validation count by New Stats, for the GL Authority, showed that MTA was undercounting Gold line "actuals" by over 10%.
By contrast, these models often underestimate highway congestion levels, and especially "variability" in congested highway speeds (models don't count accident delay, which can cause huge variance in highway travel times) when assigning trips, and often use validation for the highway congestion that may be up to 10 years out of date. So when the Gold Line extension to Azusa opened, MTA completely failed to anticipate the expected increase in ridership, overcrowding, and their assumption that half of rush hour trains would terminate at Sierra Madre Villa.
Urban travel demand forecasting, in theory, is based on good models, based on sound "scientific" observations of how folks react to changes in travel "costs" (travel time, destination and station parking cost, vehicle costs, wait time, transit fares, service reliability, etc.) and make routing and mode choices between various destinations. Unfortunately, these "forecast" models are often poorly supported by adequate data collection or monitoring of actual conditions to effectively calibrate these models, and are subject to wide variation. Changing even a small variable like wait time can have dramatic mathematical effects on outcomes, which is why those of us who use them, do so with caution.
Is SNART even running the scheduled level of service on which the 3000 daily estimate was "forecast"?
Fewer trains, longer waits, and even less frequent off peak service levels would negatively affect that "forecast" daily number. And, it has been repeatedly found that the lower levels off peak service will also cause there to be lower peak period demand (as we seen from the poor performance on the Perris line), with potential peak users who would feel "trapped" if unable to get back during the work or travel day, or after working late, or missing the last train.
I think October would be a good month to look at SMART's daily's, I did all of my major traffic and transit survey's in late September or October, or April early May, due to no holidays, no vacations, "average" day length being the same, so we would get an "normal" "average" traffic day which can be skewed dramatically by seasonal variation like length of daylight, weather, holidays, school out, beach traffic, etc..
SMART regular service response?
|
ff |
09-07-2017 - 21:13 |
Re: SMART regular service response?
|
Mike Pechner |
09-07-2017 - 21:45 |
Re: SMART regular service response?
|
GroundSwell |
09-08-2017 - 06:47 |
Re: Observations from a SMART rider
|
Dave L |
09-08-2017 - 10:43 |
Re: Observations from a Demand Forecaster |
BOB2 |
09-08-2017 - 12:30 |
Re: Observations from a Demand Forecaster
|
Dave L |
09-08-2017 - 12:53 |
SMART ridership, transit ridership, "Guaranteed Ride Home", etc.
|
Margaret_SP_fan |
09-08-2017 - 15:46 |
Re: SMART ridership, transit ridership, "Guaranteed Ride Home", etc.
|
jst3751 |
09-08-2017 - 16:16 |
Re: SMART ridership, transit ridership, "Guaranteed Ride Home", etc.
|
Ed Workman |
09-08-2017 - 17:44 |
SMART Ridership - Let's Face Facts
|
Art Carney |
09-08-2017 - 20:12 |
Re: SMART Ridership - Let's Face Facts-Yes, let's do that....
|
BOB2 |
09-09-2017 - 07:10 |
Re: SMART Ridership - Let's Face Facts-Yes, let's do that....
|
synonymouse |
09-09-2017 - 08:51 |
Re: SMART Ridership - Let's Face Facts-You just hate the vast evil train.conspiracy, we get it.......
|
BOB2 |
09-09-2017 - 09:17 |
Re: SMART ridership, transit ridership, "Guaranteed Ride Home", etc.
|
BOB2 |
09-08-2017 - 16:42 |
Re: SMART ridership, transit ridership, "Guaranteed Ride Home", etc.
|
Espee99 |
09-08-2017 - 17:16 |
Re: SMART ridership, transit ridership, "Guaranteed Ride Home", etc.
|
BOB2 |
09-08-2017 - 17:40 |
Re: SMART ridership, transit ridership, "Guaranteed Ride Home", etc.
|
espee2472 |
09-08-2017 - 18:05 |
Re: SMART ridership, transit ridership, "Guaranteed Ride Home", etc.
|
synonymouse |
09-08-2017 - 18:07 |
Re: SMART And Now for the Loony Toons Conspiracy Theory.
|
BOB2 |
09-08-2017 - 18:45 |
Re: SMART And Now for the Loony Toons Conspiracy Theory.
|
Judge Judy |
09-08-2017 - 20:58 |
Re: SMART ridership, transit ridership, "Guaranteed Ride Home", etc.
|
Erik H. |
09-09-2017 - 11:06 |
Re: Observations from a Demand Forecaster
|
Erik H. |
09-09-2017 - 11:11 |
Re: Observations from a Demand Forecaster
|
duke |
09-09-2017 - 15:01 |
Re: Observations from a Demand Forecaster
|
synonymouse |
09-09-2017 - 18:15 |
Re: Observations from a Demand Forecaster
|
Erik H. |
09-09-2017 - 19:37 |
Re: Observations from a Demand Forecaster
|
synonymouse |
09-09-2017 - 21:10 |
Re: Observations from a Demand Forecaster
|
brains |
09-11-2017 - 11:54 |
Re: Probably another member of the pro-train conspiracy
|
Just sayin' |
09-11-2017 - 12:12 |
Re: Probably another member of the pro-train conspiracy
|
espee2472 |
09-11-2017 - 13:21 |
Re: Probably another member of the pro-train conspiracy
|
synonymouse |
09-11-2017 - 15:16 |