Shared Mobility and the Transformation of Public Transit (2016)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

D-1 A P P E N D I X D Ridesourcing and Transit Travel Time Comparison

D-2 Shared Mobility and the Transformation of Public Transit While demand and capacity tell one story, another method of analysis that might be er reflect the customer experience with ridesourcing versus transit is to compare the me and cost of traveling similar routes at different mes of day. An exploratory analysis of this kind for the Chicago region compared representave travel mes for various trip types, including routes along the radial spokes of the train and highway systems, crosstown surface trips in congested corridors, circumferenal trips along the suburban periphery, and short first/last mile trips from transit terminals. The inial methodology was to simply collect the esmated driving and transit travel mes for several routes and several departure mes from the Google Maps trip planning tool, along with an esmated cost for that trip from the Uber applicaon protocol interface (API), using the UberX class of service (Figure D-1). To the driving me we added the average wait me for a pickup in Chicago, using the esmated wait me data collected during the earlier scrape of the Uber API. Results of this exploratory analysis are show below. Dollar amounts in parentheses show esmated trip cost using UberX service, from Uber API. Data sources: Google Maps (travel mes), Uber API (esmated wait mes and trip costs). Figure D-1. Representave travel mes around the Chicago region, by scheduled transit and ridesourcing, at 5:00 p.m. and 11:00 p.m. The inial analysis suggested that at peak hours, fixed-guideway trips are generally the fastest and least expensive in the corridors where they are available. At other mes of day, the marginal difference in duraon between train and ridesourcing trips would make it difficult for many riders to jusfy the much higher cost of ridesourcing based on me alone. For other trips, especially crosstown or circumferenal 0 20 40 60 80 100 120 140 Loop-MDW (radial) ($28) Loop-W. Rogers Park (radial) ($24) Wicker Park-North Ave. Beach (crosstown) ($11) Englewood-Rosemont (circumferenal) ($42) Six Corners-Montrose Blue Line (last mile) ($6) Ti m e (m in ut es ) Representave travel mes around the Chicago region TNC 5pm Transit 5pm TNC 11pm Transit 11pm Bike

Ridesourcing and Transit Travel Time Comparison D-3 ones involving mulple bus routes, the me advantage of ridesourcing was larger, and was amplified outside of peak hours when drive mes are lower and transit mes are longer. The cost poron of this analysis was based on “tradional,” single-rider ridesourcing. In cies where ride-spling versions of the service is available, ridesourcing may become more economical, and could in some places be compeve to transit in both me and cost to riders. These routes, in parcular, represent the places where transit agencies might find opportunies to shed low-ridership bus routes in favor of dynamic demand-responsive service or partnerships with shared mobility providers. For a broader analysis using the same general approach, we used another Google tool, the Distance Matrix API, which allows large-scale automated queries of their direcons system, returning a matrix of opmal travel mes for different modes from a common origin but with mulple end points. Driving mes are based on historical traffic condions for a given day and hour, and transit direcons are based on scheduled service, in each case producing an opmal route that a…empts to minimize travel me. For all of the study regions, we determined an origin address within the highest-employment census block group in each region. We programmacally queried the system for travel mes by car and by transit from the single origin to desnaon points on a half-mile grid over the core county of each study region, for both 5pm and midnight. Given the two modes’ travel mes to each point, we calculated a “travel me rao,” which is the rao of transit travel me to a derived “TNC travel me,” using the driving me figure plus average regional wait me for that hour, as obtained earlier from the Uber API (see Appendix E). Plo…ed on a map, these points give a quick overview of the tradeoffs between different modes at a regional level, as well as showing where transit is simply not an opon for a given route. Though these maps show travel between the central business district and the rest of the region, the same approach applied to a number of different origins could reveal much about the mobility picture of a given region. The maps on the following pages (Figures D-2 through D-8) show the rao of esmated transit travel me to esmated driving me (in typical traffic, plus mean TNC wait me for the departure hour and region) from a single origin to each of a 0.5 mile grid of core-county desnaons. Raos lower than 1.0 (green colors) mean that transit is faster for a given trip (the darker the green, the greater transit’s me advantage), and raos higher than 1.0 (yellow to red colors) mean that ridesourcing is faster. Points shown as only a black dot represent areas for which no rao could be calculated because either a) no transit route exists between the origin and desnaon; or b) they represent points with no public roads, such as airports, gated subdivisions or undeveloped areas. While the specific findings emerging from this analycal approach vary from city to city, a few pa…erns emerge: • Peak hour traffic congeson ps the scales in favor of transit that travels in its own right of way--on tracks above or below traffic, or in dedicated lanes. • Long transit headways at night, along with easier travel on largely congeson-free streets, mean that TNCs are the faster mode for many desnaons; but cost remains a key determinant of whether this is actually a viable choice for frequent trips.

D-4 Shared Mobility and the Transformation of Public Transit • In a few places (central Ausn and Seale, for instance), certain trips are faster on transit late at night than at rush hour--reflecng how congeson blocks the effecveness of transit running in mixed traffic. Also note that the maps do not account for the differing cost of rides on transit versus TNCs. As distances increase--and costs with them--it is likely that for most users, the appeal of even a relavely faster TNC ride would drop significantly beyond a certain cost threshold. For occasional trips this might not be a central consideraon, but for more frequent trips these costs would be unsustainable. For many trips in these areas, the personal automobile is likely to remain as the mode that maximizes ulity for the individual traveler, unl some combinaon arises of a) wider coverage of more frequent transit or b) much lower cost TNC services, such as shared ride services. Figures D-2 through D-8 combine data from Google Maps Distance Matrix API (transit and driving me esmates), Uber API (TNC wait me), and US Census Bureau TIGER/Line (geography).

Figure D-2. Transit-TNC travel me rao, 5:00 p.m. and midnight, Ausn, TX.

Figure D-3. Transit-TNC travel me rao, 5:00 p.m. and midnight, Boston, MA.

Figure D-4. Transit-TNC travel me rao, 5:00 p.m. and midnight, Chicago, IL. Cook Cty grid-0.5 mi Central city Origin point 1 10 3 5 mi.2 4 Transit:TNC
me ra
o 2.00 or more Transit:TNC
me ra
o 1.50-2.00 Transit:TNC
me ra
o 1.00-1.50 Transit:TNC
me ra
o 0.50-1.00 Transit:TNC
me ra
o 0.00-0.50

Figure D-5. Transit-TNC travel me rao, 5:00 p.m. and midnight, Los Angeles, CA.

Figure D-6. Transit-TNC travel me rao, 5:00 p.m. and midnight, San Francisco, CA.

Figure D-7. Transit-TNC travel me rao, 5:00 p.m. and midnight, Seale, WA.

Figure D-8. Transit-TNC travel me rao, 5:00 p.m. and midnight, Washington, DC.