Scenario Comparison
| Scenario | Strategy | Intervention | Tracts Affected | Avg. System Impact |
|---|---|---|---|---|
| S1 ★ | Blanket Bus Expansion | +20% bus stops, +20% density, −20% dist_to_bus — all tracts | All 2,165 | +0.61 pp avg. |
| S2 | Targeted Bus Investment | +50% bus stops/density, −40% dist_to_bus — 215 high-need tracts only | 215 (9.9%) | +0.15 pp in targets |
| S3 | Targeted Rail Expansion | +1 rail station, −40% dist_to_rail — top-quartile predicted tracts | 541 (25%) | Mixed in targets |
Note on Scenarios
These are not forecasts. They are sensitivity tests that estimate how ridership would respond to changes in transit access, based on relationships the model learned from current data. They show the relative impact of different investment strategies — not a predicted future state.
Scenario Detail
Scenario 01 · Blanket Expansion
Uniform Bus Infrastructure Increase (+20%)
Bus stops increased by 20% system-wide, bus density scaled proportionally, and average distance to the nearest stop reduced by 20% across all 2,181 tracts. This represents the default planning approach: grow the network uniformly and let ridership follow.
The resulting map shows modest gains distributed broadly across the region. The largest absolute impacts appear in areas where baseline service is already reasonable — a predictable pattern, since well-served tracts sit on the steeper part of the density-ridership curve. Low-access tracts improve, but not enough to cross the 1-mile threshold where ridership meaningfully increases.
+0.61 pp avg. system-wide · 1,110 of 2,165 tracts improved ★Predicted change in transit share — Scenario 1 vs. baseline · Uniform +20% bus infrastructure
Scenario 02 · Recommended Strategy
Targeted Bus Investment in High-Demand, Low-Access Tracts
Bus improvements are concentrated in tracts that are simultaneously in the top quartile of a composite demand score (weighted by population, pct_hispanic, and pct_foreign_born) and in the bottom half of existing bus density. In target tracts only: bus stops +50%, density +50%, distance to nearest stop −40%.
This scenario concentrates investment in 215 high-need tracts — the communities with the clearest gap between demand and access. The absolute system-wide gain is smaller than Scenario 1, but the reach is equity-focused: every investment unit goes toward the tracts where unmet demand is greatest. This is what gap analysis enables — a spatial priority queue for investment targeting rather than uniform expansion.
+0.15 pp avg. in 215 targeted tracts · equity-focused reachPredicted change in transit share — Scenario 2 vs. baseline · Targeted bus in high-gap tracts
Scenario 03 · Rail Expansion
Targeted Rail Infrastructure in High-Responsiveness Corridors
Rail stations are added (+1 station) and distance to rail is reduced (−40%) in tracts in the top quartile of baseline predicted transit share — areas where the model already identifies strong ridership potential and where rail investment is most likely to trigger a step-change in mode choice.
Gains are concentrated in a subset of high-responsiveness corridors. The map reveals a pattern consistent with the overall findings: rail expansion alone, without addressing the density of the bus network that feeds rail stations, produces corridor-specific gains with limited geographic reach. The greatest impact comes from combining Scenario 3 corridors with the Scenario 2 bus investment strategy — a hybrid approach.
Mixed results · already well-served areas show limited model responsivenessPredicted change in transit share — Scenario 3 vs. baseline · Rail expansion in top-predicted corridors
Planning Implication
The Tradeoff
Blanket expansion wins on volume. Targeted investment wins on equity.
Scenario 1 produces more total ridership gain. Scenario 2 reaches the communities with the greatest unmet need. The gap map is the decision tool: it translates model output into a spatial priority queue that can inform either strategy.
The Hybrid Strategy
Bus first, rail to complement.
Rail expansion (Scenario 3) produces the most impact in corridors where bus density already brings riders to stations. The highest-return strategy is closing the bus network gaps first (Scenario 2), then adding rail in the corridors where that network already performs.
Portfolio Signal
The model is a planning instrument, not just a prediction.
These scenarios are not forecasts — they are sensitivity analyses. By perturbing infrastructure inputs and re-running the Random Forest, we translate the model's learned relationships into investment priorities. The approach demonstrates that data science and spatial analysis can do more than describe current conditions: they can simulate policy decisions and rank alternatives before a dollar is spent.