The LRTP Prioritization Problem
Metropolitan Planning Organizations are responsible for developing Long Range Transportation Plans that identify and prioritize transportation investments over a 20–25 year horizon. Federal regulations under 23 CFR Part 450 require that LRTPs be fiscally constrained — project lists must reflect realistic revenue projections, not wish lists. That constraint means prioritization decisions matter: with a finite federal aid envelope and limited state and local match capacity, MPOs have to make choices about which corridors get BRT investment, which arterials get signal priority upgrades, which transit extensions get funded first.
In practice, those prioritization decisions are often made using a combination of travel demand model outputs, staff judgment, and political negotiation among member jurisdictions. The travel demand model outputs — projected link volumes, mode split estimates, travel time savings — provide quantitative foundation for the analysis. But those models have limitations that planners understand intuitively but often struggle to surface explicitly in the public project prioritization process.
The core limitation: regional travel demand models are calibrated against historical behavior and project future behavior using relatively coarse spatial and temporal assumptions. For corridor screening — identifying where demand is concentrated right now and where it's growing — these models tend to be less useful than current ridership and mobility data analyzed at the resolution of the actual travel pattern. A demand surface built from GTFS ridership data, GBFS bike-share activity, LEHD employment data, and ACS demographic data gives MPO planners a current-conditions picture that travel demand models, by construction, cannot provide.
What "Data-Driven Prioritization" Actually Means
The phrase gets used loosely, so it's worth being precise. A data-driven LRTP corridor prioritization process means, specifically:
- Screening candidate corridors against current observed demand data — not just projected future demand — to identify where investment will serve an already-demonstrated travel pattern vs. where it's betting on induced demand or land use change.
- Quantifying equity dimensions of the prioritization using demographic data at the spatial resolution of the corridor analysis — not just checking whether the corridor passes through a census tract with above-average minority population.
- Documenting the analytical basis for ranking decisions in a way that's reproducible and auditable — allowing corridor rankings to be revisited as conditions change without requiring a complete re-analysis.
- Being explicit about uncertainty — acknowledging where demand projections are well-grounded (based on current observed patterns) and where they're speculative (based on assumed land use change or behavior shift).
This is a higher analytical standard than most current LRTP processes meet. We're not saying MPOs are cutting corners — the current standard reflects real resource constraints and the methodological state of the practice. What we're arguing is that current-conditions demand surface data substantially reduces the analytical cost of meeting this higher standard, if it's integrated into the LRTP process appropriately.
The Role of a Demand Surface in LRTP Corridor Screening
The specific analytical contribution of a demand surface in the LRTP process is at the corridor screening stage — the phase where a long list of potential investments gets filtered down to a shorter list for detailed analysis. At this stage, the key question is: where does concentrated, unmet, or underserved travel demand exist in the region right now?
A demand surface built on H3 resolution 7 cells (appropriate for regional-scale MPO analysis) and covering 12–24 months of ridership data can answer that question spatially with a resolution that travel demand model outputs typically cannot. Specifically:
- Identifying corridors where current ridership demand is concentrated but service is inadequate: High demand density in H3 cells that are currently served only by low-frequency routes, or not served at all, is a strong signal for investment prioritization. This is observable in current data, not dependent on a demand model projection.
- Identifying corridors where projected land use change will amplify existing demand: Overlaying demand surface data with regional land use plans and development pipeline information lets analysts identify corridors where current demand is already significant and approved development will further concentrate demand — a higher-confidence investment case than pure greenfield demand projection.
- Equity-weighted prioritization: H3-level demand data overlaid with ACS demographic data allows analysts to explicitly weight corridors by the proportion of low-income, car-free, or transit-dependent population served — producing an equity-adjusted prioritization ranking that can be documented and defended in the public process.
A Scenario: Four-County MPO Corridor Ranking
Consider a four-county MPO in a growing mountain west region with a LRTP update cycle underway. The MPO has 14 candidate transit corridor projects in the project universe — a mix of BRT conversions, frequency enhancements, and new service extensions. The standard approach would be to run each corridor through the regional travel demand model and rank by projected boardings, cost-effectiveness (cost per boarding), and GHG benefit.
Adding a demand surface analysis to that process produces a different kind of input. The demand surface identifies three corridors where current ridership density in H3 cells is substantially higher than the travel demand model projects — corridors where the model is underestimating existing demand, likely because it's calibrated on older behavior data. It also identifies two corridors where the model shows high projected boardings but current observed demand is low — corridors where the model is relying on speculative land use change assumptions that may not materialize.
This doesn't override the travel demand model results; it contextualizes them. The MPO planning staff can now present the corridor ranking with a demand-confidence qualifier: "high observed demand / model confirms" vs. "model-projected demand / limited current ridership signal" vs. "strong current demand / model may be understating." That distinction matters for how the MPO presents projects to federal funding agencies and for how it structures the financial plan around projects with different demand certainty profiles.
Connecting Demand Analysis to CMAQ Eligibility
The Congestion Mitigation and Air Quality Improvement (CMAQ) program provides federal funding for transportation projects that reduce congestion and improve air quality in non-attainment and maintenance areas. CMAQ project eligibility requires demonstrating VMT reduction — specifically, that the project will reduce vehicle miles traveled by diverting trips from automobile to transit or other non-single-occupant vehicle modes.
A demand surface analysis can directly support CMAQ justifications by providing observed mode shift data: where multimodal demand data shows trips that are currently made by bicycle, transit, and micro-mobility in corridors with significant auto congestion, those observed multimodal trip patterns provide a documented baseline against which VMT reduction estimates can be calibrated. This is analytically stronger than relying entirely on travel demand model mode split projections.
The documentation trail from observed multimodal demand data to CMAQ VMT reduction estimate is more defensible in federal review than model-only projections, particularly now that travel demand models are known to have calibration uncertainty from the pandemic-era disruption to observed behavior data.
Making the Analytical Process Defensible
LRTP processes are inherently political — member jurisdictions advocate for projects that serve their communities, and the planning process has to manage those competing interests transparently. A data-driven prioritization process doesn't eliminate that political dimension, but it changes the character of the conversation. When a corridor's prioritization ranking is grounded in observable demand data with documented methodology, the conversation shifts from "why wasn't our project ranked higher?" to "here's the data, let's discuss whether the methodology is capturing the right demand factors." That's a more productive conversation, and it's more defensible in federal planning certification reviews.
We're not saying data-driven analysis resolves political trade-offs in LRTP development. It doesn't. What it provides is a shared analytical foundation that makes the trade-offs visible and the prioritization decisions explainable to the public, to member jurisdictions, and to federal oversight agencies.