This repository contains some R code and data files related to simulating passenger loads and schedules on the MBTA Framingham/Worcester Line. There are significant caveats to this; please read my blog post for more information about the methodology and data sources. Patches and improvements gladly accepted.
The following changes have been made since the initial blog post:
- Use of global variables has been substantially reduced, to allow for the model to be generalized and eventually applied to other lines and other scenarios.
- Other places where values were hardcoded have been replaced with code to deduce or generate the required data.
- We now run 250 Monte Carlo trials and output the 90th percentile loadings rather than just running the simulation once and outputting the exact simulated loadings from that one run.
- The unit requirements are computed, based on the results of the simulation and a parameter which is the number of passengers per unit.
- Improvements have been made to the schedule calculations, in preparation for adding the ability to handle more complex service patterns like short turns and expresses.
The files in this repository are as follows:
-
2012-boardings.csv: net inbound Framingham/Worcester boardings at all stations for all AM departures, taken from the 2012 CTPS ridership audit ** NOTE WELL: CTPS is presently conducting a ridership study for 2018 to assist in MassDOT's planning process. -
Framingham-Worcester Inbound.{numbers,pdf}: the spreadsheet I used to work out the schedule of various local and express configurations as well as equipment and storage requirements, with a viewable/printable PDF if you just want to look at the results -
emu-schedule.csv: a schedule for an inbound Regional Rail local train, based on Alon Levy's modeling for North-South Rail Link, with Alon's proposed infill stations removed -
train.sched.r: some very bad R code for simulating passenger counts based on observed boardings; uses lots of global variables and needs to be refactored. See the blog post for additional caveats. -
YYYY-MM-DD: output files generated by various versions of this script at the indicated date ** The scenarios modeled are generally different for each run.