Why
In the Phoenix of 2024, only rail lines attract any choice riders,1 and those riders are generally traveling short distances, say between Apache Blvd and ASU or Midtown and Downtown Phoenix. The first step to a regional transit system for everyone is faster rail.
In a Phoenix where transit between Phoenix and Tempe is time competitive with driving, more people choose to take rail, buses feeding the light rail generate more ridership and justify higher frequencies, and the region’s support for transit increases.
What’s next for rail?
after i10, future projects in phoenix will be relatively slow extensions of already relatively slow lines (and Phoenix knows it – they’ve been paused)
this creates a rapid transit backbone that’s good for existing and planned extensions, but also encourages cities (incl cities who haven’t liked rail in the past) to think about extensions that connect them (no neighborhood in Seattle is against Link Light Rail because it’s fast, and therefore for everyone)
Transit is only competitive when it’s time efficient compared to its alternatives.
[Add sources].
It’s at a disadvantage, especially in a city like Phoenix.
[Add why cars are fast and transit is slow]
[Link TSP or this page from mentions of TSP in Trunk page]
Commercial TOD
Specific pages for trunk and TSP and downtown tunnel maybe?
- The choice vs captive rider dichotomy is a fraught oversimplification which should be thought of as a spectrum, but it’s useful in the most general sense to understand how it becomes socially acceptable for anyone of any income and social class to ride transit in a metro area. ↩︎
Add a link to this page from the Trunk Felt Map “Univ and Rural Grade Separation” blurb
Add link to TSP page from “average station dwell data collected Fall 2024” in Trunk footnotes
Other pages to make:
- Creating track generator with ChatGPT
- what ChatGPT is good vs bad at: will bring up some novel concepts but often it will interpret you too literally and not do a much easier way of coding; doesn’t know how to make them work with adding/subtracting; not up to date or correct on APIs; good with general code structure; me and it are bad at making code truly good; you have to break down the problem for it; they don’t understand how QGIS works (geometry generator vs default expression vs python window); being really specific is important (and you don’t want to not specify everything and get stuck in path dependency) but also specifying everything in ChatGPT means less work for me later (put the layer names in to ChatGPT so I don’t have to manually add them in QGIS); it will hallucinate methods; often it’s best to ask it to do it at a high level to see how it chooses to do it (often something I haven’t thought of), then if it doesn’t implement it correctly, explain step by step how to do that thing for it to turn that into code
- Gathering data for VM Stats and analyzing (mention and link to above about Chat GPT for code but don’t go in detail)
- Tempe Enhanced TSP Pilot