Frequency Vs Coverage Model

The city has stated that they are moving from a coverage model to a frequency model. This is evident in the following summary:

System Active Stops Weekly Departures Ave. Departures Per Stop Ave. Dist. (m) Closest Stop (Euclidian) Ave. Dist. (m) Closest Stop (Grid)
June 1465 261,868 178.7 270.5 331.7
July 1443 275,142 190.7 269.4 330.3
BRT 994 217,543 218.8 351.7 425.4

The average departures per stop are generated from plot_routes.py, whereas the average closets distance is generated via a heatmap with a certain mode.

As can be seen, the number of active stops in the BRT system has been reduced substantially. As expected, this has caused the average distance to the closest active stop has increased. The total number of weekely departures in the BRT system declined substantially, but the departures per stop have increased.

These results illustrate the move from the coverage based model to a frequency based model.

The average distance to the closest stop was computed as follows:

  • The DAs are divided into 100m X 100m cells.
  • A set of DA cells to be used in all subsequent analyses are determined by discarding any DA cell more than 1000m away from the closest stop in the June dataset. This removes many DA cells from the analysis, especially in the DAs around the peripheries of the city. These are shown in a plot elsewhere on this wiki (inset link here). These are called the active DA cells.
  • For all the active DA cells, determine the distance to the closest active stop (i.e., a stop that serve at least one route), using either a Euclidian measure or a Grid measure.
  • Compute the average distance for all the active DA cells.
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License