Sharpening The Transportation Network Connectivity Metrics

Andy Mortensen, Consultant at Transpo Group

The 2009 CNU Transportation Summit was a great success. Thank you Summit organizers, CNU staff and presenters.

There was good discussion and work towards refining CNU’s Statement of Principles on Transportation Networks, most notably this excerpt from the Preamble:

“a network-level approach to transportation planning and design recognizes that a transportation system’s primary purpose is to connect people, places and activities with each other and with the social and economic activities of the larger community and region…”

It is about connectivity, and I think most agree further that “good street networks are highly connected.” Connectivity was either the direct or at least indirect theme in all of the presentations and sessions. We heard reporting on different applications to measure transportation connectivity by use of intersection density scoring, Walk Score™, and the Route Directness Index (RDI).

Intersection density and Walk Score™ maps are visually appealing and effective when comparing disparate connectivity from one subarea or city to another. I will argue they lack real analytical power and merit, especially at the neighborhood level, where we win or lose the connectivity struggle. The RDI, on the other hand, is a direct measure of connectivity and yields quantified scoring at the block level of geography.

What is RDI? RDI is calculated by dividing direct (“crow flight”) travel distances by actual travel distances on a scale of 0.1 – 1.0. For example, if streets are connected, patterned in small blocks, and have good sidewalks, people can travel nearly directly to most destinations thereby resulting in a higher index value. If the street network has many unconnected dead-ends and blocks are large, people must travel farther to reach destinations, resulting in a lower index value. An RDI of 1.0 is the best possible rating, indicating that pedestrians can walk directly to all destinations. A score of 1.0 is not likely achievable in the best of connected land use settings; an average value of 0.60-0.70 is considered by some as acceptable. Opinions will certainly vary by city and even within urban areas. If you’re interested in RDI, you can see the CNU presentation run here:

As the truest measure of network design (opinion), RDI-scored network connectivity, combined with land diversity and density measures, enables planners and engineers to complete their geospatial analyses and prioritize transportation solutions that are likely to yield the greatest benefit in improved connectivity - ultimately helping achieve the broader goals of reduced VMT and GHG per capita. If developed and applied appropriately, the RDI will also become the platform by which local multi-modal quality of service (QOS) measures will be used to supplement existing concurrency and growth management policies, which have been dominated by vehicle level of service measures. Many cities are now seeking alternate QOS measures, particularly for pedestrian and bicycle modes, and are frustrated in the lack of progress towards their development. RDI is part of the solution.

I wish I caught the person’s name who spoke during the open-mike session last Friday afternoon, and sorry if I’m not quoting you perfectly here, but I think he said: “We’ve already done the general comparison studies; it’s time we get to the detailed measures.” I agree, and will take the risk of stating it in perhaps stronger terms: at the local level, we need to focus on actual and direct measures of connectivity, not relative or proxy measures.

Of course I’m partial to RDI, and gave a presentation at the CNU Summit (see link above) summarizing its merits and comparing systematized RDI to other proxy measures of connectivity. I think the RDI measure first trains the eye more specifically to the real connectivity gaps in our transportation networks. It then provides a meaningful, numeric baseline measure by which we can gauge the level of connectedness in our individual transportation systems, from which we can more systematically identify and prioritize improvements (maybe refine our plans ?). Systematic application of RDI is particularly helpful in transit station area and walk-to-school routing analyses.

The City of Portland presented some of their plan findings at the Summit and noted the significant and unmet funding needs to correct connectivity problems, particularly near LRT station areas. In one form or another, these issues reside in all of our cities. Intersection density mapping helps target the general problem, but provides no further assessment mechanism to identify and prioritize projects on their merits of improving connectivity. If federal and state funding is to be more focused on improving connectivity, and the dollars are likely to be scarce and the competition steep, then there should be some commonality in connectivity measurements, at the local level, in each city and state.

I’m not short on opinion. I think RDI measurements should be used to establish baseline policy and connectivity thresholds by land area type, as a quality of service measure.

Suppose that any new federal investment to local transportation solutions is allocated to jurisdictions only when they demonstrate measured contribution towards improved connectivity? Let’s further suppose that the measurement of connectivity be based on the following minimum RDI thresholds (as a pre-qualification for a return on investment):

RDI Score
Downtown / Centers > 0.75
Transit Stations > 0.75
Walk-To-School Zones > 0.75
General Urban > 0.70
General Suburban > 0.65

This may be oversimplified, but it’s a starting point for discussion. Should each community examine their systems and establish their own policies? Definitely, but exceptions should not be made for poorly connected, urban and suburban street patterns. There is merit in establishing national guidelines and complex formulas may not be the answer. Look at the Highway Capacity Manual, in its current form and the forthcoming update. Lot’s of numbers, equations and technical recommendations, mostly focused on the automobile - after all, it is the Highway Capacity Manual – but now with a new banner for “multi-modal.” At the risk of alienating some traffic engineer friends, with all its details, look where heavily reliance on the HCM’s application has gotten us over the past decades. From a multi-modal perspective…..are we safer? Do we travel (not just drive) more efficiently? Have we spent our public dollars wisely on the range of transportation solutions? Have we adequately planned for future generations?

I will say one thing: the HCM, when it’s used extensively and singularly to establish and apply system performance tools and policies, is not providing adequate direction to plan and build connected street networks - networks that are used by the range of walkers, cyclists, drivers and passengers.

Connectivity is what it’s all about. New direction is needed. Kindly share your thoughts.

Andy Mortensen
Transpo Group


Write your comments in the box below and share on your Facebook!