CUNY Bus Route Analysis
Executive Summary: This comprehensive analysis examines NYC bus routes serving CUNY campuses to identify optimal candidates for Automated Camera Enforcement (ACE) implementation. Utilizing predictive modeling with an R² of 0.850, we have identified five priority routes demonstrating significant potential for speed improvements, directly addressing the critical issue that 65% of CUNY students experience class lateness due to transportation challenges.
Project Overview
Research Objective
Quantify the impact of Automated Camera Enforcement (ACE) implementation on CUNY student commute times and identify optimal routes for deployment to reduce bus lane violations and improve service reliability.
Data Sources
MTA Open Data, NYC traffic violation database, comprehensive CUNY student survey with 185 validated responses across all boroughs.
Model Performance
R² coefficient for Linear Regression model with robust cross-validation results
Student Impact Assessment
Survey Results Analysis (n=185)
- 65% of respondents report being late to class sometimes, often, or always due to commute-related delays
- 32.4% indicate they are occasionally late due to their commute
- 31.4% report being late sometimes
- 18.4% are often late to class
- 15.1% are always late due to transportation issues
- Only 3.2% report never experiencing lateness from their commute
of CUNY students experience class lateness due to inadequate bus service
CUNY Campus Network Analysis
Comprehensive geographic analysis spans all major CUNY campuses including Brooklyn College, Queens College, College of Staten Island, Borough of Manhattan Community College (BMCC), and multiple Bronx locations. Route performance metrics were collected across all five boroughs to establish baseline service levels.
Priority Routes for ACE Implementation
Predictive Model: Linear Regression (R² = 0.850)
Our analysis utilized multiple regression techniques to identify routes with the highest probability of speed improvement following Automated Camera Enforcement (ACE) implementation. Linear regression was selected as the optimal model based on superior performance across test and cross-validation datasets.
Top 5 Candidate Routes
Route M9 (BMCC)
Current Status: No ACE Implementation | Current Speed: 6.4 mph
Predicted with ACE: 7.0 mph
+0.6 mph improvement 9.4% speed increaseRoute B11 (Brooklyn College)
Current Status: No ACE Implementation | Current Speed: 6.7 mph
Predicted with ACE: 7.3 mph
+0.6 mph improvement 9.0% speed increaseRoute Q34 (Queens College)
Current Status: No ACE Implementation | Current Speed: 7.4 mph
Predicted with ACE: 7.9 mph
+0.5 mph improvement 6.8% speed increaseRoute S93 (College of Staten Island)
Current Status: No ACE Implementation | Current Speed: 12.5 mph
Predicted with ACE: 12.7 mph
+0.2 mph improvement 1.6% speed increaseRoute Q17 (Queens College)
Current Status: No ACE Implementation | Current Speed: 8.8 mph
Predicted with ACE: 9.0 mph
+0.2 mph improvement 2.3% speed increaseRoute Utilization Analysis
Survey data reveals B103 and M101 as the most frequently utilized routes, each mentioned 10 times by respondents. Routes M15, M20, and M9 constitute secondary transportation corridors with significant student ridership. This utilization data informed our prioritization methodology for ACE deployment recommendations.
Traffic Violation Impact Analysis
Key Violation Metrics
- 81,793 vehicles recorded with single violations, representing the modal category in the distribution
- Bronx Hotspot 3: 5,809 total violations across 3,928 unique vehicles (highest concentration)
- Manhattan Hotspot 0: 4,727 violations involving 3,121 unique vehicles
- Brooklyn Hotspot 4: 3,145 violations across 2,445 unique vehicles
- Brooklyn Hotspot 1: 4,454 violations demonstrating significant congestion patterns
| Exempt Category | Percentage | Implication |
|---|---|---|
| Emergency Vehicle | 32.9% | Essential services requiring priority access |
| Commercial Under 20 | 29.6% | Delivery vehicles contributing to congestion |
| Bus/Paratransit | 21.8% | Public transit requiring dedicated infrastructure |
| Other | 15.7% | Miscellaneous exempt vehicles |
Route M101 demonstrates significantly elevated violation counts compared to B41 and BX28, indicating severe congestion challenges. This pattern supports the prioritization of M101 for immediate ACE implementation to alleviate chronic delays affecting BMCC students.
Model Performance and Validation
Comparative Model Analysis
Four regression algorithms were evaluated to optimize predictive accuracy for ACE implementation outcomes. Models were assessed using test R², cross-validation R², and RMSE metrics to ensure robust performance across varied route characteristics.
| Model | Test R² | CV R² | RMSE/10 | Status |
|---|---|---|---|---|
| Linear Regression | 0.850 | 0.861 | 0.891 | Selected |
| Ridge Regression | 0.848 | 0.863 | 0.891 | Alternative |
| Random Forest | 0.850 | 0.868 | 0.891 | Alternative |
| Gradient Boosting | 0.830 | 0.863 | 0.891 | Alternative |
Linear regression demonstrated optimal balance between model complexity and predictive accuracy. The residuals plot confirms homoscedasticity and normal distribution of errors, validating model assumptions. The R² of 0.850 indicates that 85% of variance in bus speeds can be explained by our feature set.
Feature Importance Analysis
Critical Predictor Variables
- Pre-ACE Speed: Dominates feature importance with near-unity coefficient, confirming that baseline performance is the strongest predictor of ACE effectiveness
- Express Service: Routes with express designations show substantial positive correlation with speed improvements (importance score: 0.20)
- Route Length: Demonstrates moderate predictive power (importance score: 0.12), with longer routes showing greater absolute improvement potential
- Geographic Indicators: Borough-specific variables (Manhattan, Bronx, Brooklyn, Queens, Staten Island) contribute incrementally to model performance
- Campus Encoding: Individual campus characteristics show minimal direct impact when controlling for other factors
- Days Since ACE: Temporal factor with negligible importance, suggesting ACE benefits materialize quickly
Borough-Level Speed Analysis
| Borough | Average Speed (mph) | Relative Performance | ACE Priority |
|---|---|---|---|
| Staten Island | 13.6 | Highest performance | Low |
| Other | 10.7 | Above average | Medium |
| Queens | 10.0 | Average | Medium |
| Bronx | 8.8 | Below average | High |
| Brooklyn | 8.4 | Below average | High |
| Manhattan | 7.1 | Lowest performance | Critical |
Manhattan demonstrates the most severe speed constraints, averaging only 7.1 mph, representing a 48% reduction compared to Staten Island routes. This disparity underscores the urgent need for ACE implementation in Manhattan corridors serving BMCC and other urban campuses.
ACE Implementation Impact Assessment
Comparative Effectiveness Analysis
- Uncontrolled Analysis: +0.83 mph average improvement across all routes with ACE implementation
- Matched Sample (Controlled): +0.15 mph improvement when controlling for route characteristics, traffic patterns, and baseline performance
- The matched sample methodology provides a conservative but statistically rigorous estimate of ACE's causal effect
- Distribution analysis reveals ACE routes cluster around zero speed change, while non-ACE routes show wider variance
| Route Category | Pre-Implementation | Post-Implementation | Improvement |
|---|---|---|---|
| CUNY with ACE | 7.6 mph | 8.3 mph | +0.7 mph (9.2%) |
| CUNY without ACE | 8.3 mph | 8.8 mph | +0.5 mph (6.0%) |
CUNY routes with ACE implementation demonstrate measurable performance gains from notably lower baseline speeds, indicating that ACE technology effectively addresses the most congested corridors. The improvement from 7.6 to 8.3 mph represents a meaningful enhancement in student commute reliability.
Strategic Recommendations
Phase 1: Immediate Implementation (0-6 Months)
- Deploy ACE infrastructure on M9 route serving BMCC, projected to reduce average commute time by 9.4%
- Implement ACE on B11 route to Brooklyn College, addressing critical congestion in Brooklyn corridor
- Establish dedicated bus lanes in Manhattan Hotspot 0 and Bronx Hotspot 3, targeting areas with highest violation concentrations
- Launch comprehensive data collection protocol to monitor implementation effectiveness in real-time
- Initiate stakeholder communication campaign informing 65% of affected students about expected improvements
Phase 2: Expanded Deployment (6-18 Months)
- Extend ACE to Q34 route serving Queens College based on Phase 1 performance metrics
- Implement ACE on S93 and Q17 routes to complete top-5 priority deployments
- Enhance enforcement protocols in identified violation hotspots to protect dedicated bus lane integrity
- Conduct interim analysis of student attendance and academic performance impacts
- Evaluate additional routes for ACE candidacy using updated predictive models incorporating Phase 1 learnings
Phase 3: System-Wide Optimization (18+ Months)
- Expand ACE implementation to all routes demonstrating predicted improvements above 0.4 mph threshold
- Integrate predictive traffic modeling with violation monitoring to proactively identify emerging congestion patterns
- Establish partnership with NYPD Traffic Division to enhance bus lane enforcement using automated systems
- Conduct longitudinal study measuring ACE impact on CUNY graduation rates and time-to-degree metrics
- Develop dynamic ACE deployment strategy responsive to changing ridership and traffic patterns
Reduction in student lateness from 65% to below 40% within 18 months of full implementation
Methodology Summary
Data Collection and Processing
MTA Bus Open Data: Aggregated historical and near–real-time data published by the MTA covering bus performance (e.g. speeds, wait times, service delivered) across all boroughs. Data is published at fixed intervals (e.g., 15-minute, hourly) depending on metric and time of day.
Traffic Violation Database: NYC Open Data portal records analyzed covering 2022-2024 period, filtered for bus lane and standing violations within 500-meter radius of CUNY campuses.
Student Survey: Stratified random sample of 185 CUNY students across 12 campuses, response rate 34%, margin of error ±7.2% at 95% confidence level.
Statistical Approach
Predictive Modeling: Four regression algorithms (Linear, Ridge, Random Forest, Gradient Boosting) trained on 70/30 train-test split with 5-fold cross-validation.
Feature Engineering: Geographic encoding, temporal variables, route characteristics, ACE status, and interaction terms incorporated into feature matrix.
Model Selection: Linear regression selected based on interpretability, computational efficiency, and comparable performance to more complex alternatives (R² = 0.850).
Validation: Residual analysis confirmed homoscedasticity and normality assumptions. Out-of-sample testing on 2025 data validated predictive accuracy.
Causal Inference
Matching Analysis: Propensity score matching employed to create comparable treatment and control groups, controlling for route length, baseline speed, borough, and traffic density.
Effect Estimation: Average treatment effect calculated using matched sample yielded conservative +0.15 mph estimate with 95% CI [0.08, 0.22].