Mystery models of the Northern Branch

I'm old enough to remember the days when computers were expensive enough that the only people who could afford one had either a lot of money or an IBM employee discount. Even when it got to the point that most households had a computer, there were still a lot of programs that were so complex they needed a large mainframe to run. Then there were proprietary programs that required expensive licenses, and the number of trained programmers was relatively small. Many governments, even if they wanted to keep their business open to the public, had no choice but to restrict access to their programs.

Those days are mostly over. Today, my cell phone can do a lot more than the average mainframe could in 1982. A lot of software is open source and can be downloaded and shared for free. A relatively inexperienced programmer can may not be able to delve into the depths of some programs, but can still take an existing program and create a freely accessible interface for it on the World Wide Web.

In short, in this day and age there is really no reason for a government to keep their computer programs secret. Any code that was created by government employees should be open source, and ideally should be able to be run by any citizen. Applications that analyze large databases can be opened up with a Web interface and an internet application-programmer interface so that independent programmers can connect to them.

The proposed site of the North Tenafly station. 8.6 acres (0.68 Jakriborgs) of space, including everything across the street, would be leveled to make room for a 570-space parking lot.

Unfortunately, transportation modeling has not moved into the twenty-first century. Case in point: the Northern Branch Corridor Project in New Jersey, whose managers would "prefer" to plop 2,310 parking spaces up and down the train line. I combed through the Draft Environmental Impact Statement - it's pretty readable as these things go - looking for New Jersey Transit's justification for building that many spaces, and the best I could find was this section from the Executive Summary (PDF):

ridership for each of the Build Alternatives was projected to the year 2030 using the North Jersey Transit Demand Forecasting Model (NJTDFM, the model). The model was used to estimate total LRT riders by zone to each station. NJ TRANSIT then used data from the model (such as percentage of people that would walk, drive, carpool, be dropped off, or take the bus based upon distance from each station, and the type and density of development) to estimate how riders would travel to each station. This produced an initial parking demand at each station. The station locations were then reviewed to determine the maximum amount of parking spaces that could reasonably be provided at each location, without overwhelming the communities. The model was then constrained with these maximum parking numbers, and a second iteration of the forecasting model was run. This resulted in adjustments to individual access modes to each station and a final, revised parking demand at each station

So, why build that much parking? Because our computer model says that there will be that much demand. Why does the computer model say that? Does it take into account the effect of parking lot capacity and price on the demand for parking? Does it account for the effect of zoning changes (for example, to allow dense development without accessory parking adjacent to the stations) on parking demand and drop-offs? Who knows.

I emailed one of the senior transportation demand forecasters at New Jersey Transit last week, and asked him these questions. I also asked about public access to the NJTDFM. If I hear back, I'll let you know.