Inside the sports complex
Edge-AI cameras reveal when and where fans arrive and depart across the NJ TRANSIT Rail Area, Pedestrian Bridge, and NJ TRANSIT shuttle-bus terminal.
How do tens of thousands of fans move through the Meadowlands before and after a World Cup match? This dashboard, developed through research by the NYU C2SMART Center, combines unique Edge-AI observations inside the sports complex with third-party roadway data. Findings cover multiple World Cup matchdays at MetLife Stadium, including the July 19 final.
Edge-AI cameras reveal when and where fans arrive and depart across the NJ TRANSIT Rail Area, Pedestrian Bridge, and NJ TRANSIT shuttle-bus terminal.
Traffic analysis is being prepared for a future public release.
Cross-match views compare weekend afternoons, weekday afternoons, and weekday evenings before users explore an individual match.
Explore changes in pedestrian flow and peak bottleneck intensity following construction of the new pedestrian bridge and coordinated World Cup crowd-management operations.
Pedestrian insights are preliminary and are derived from an enhanced computer-vision crowd detection method developed by the NYU C2SMART team using two Edge AI camera views. The estimates reflect pedestrian activity visible within the monitored fields of view and counting lines and do not represent total attendance or all pedestrian movements in and around the venue.
Privacy and Data Protection Note: Video data are used solely to analyze aggregate pedestrian and transportation patterns. Any imagery used in the research is anonymized so that individuals' faces are not recognizable, and the analysis does not identify or track specific people. The dashboard presents only de-identified, aggregated counts and movement patterns.
Although the detection method has been enhanced for high-density crowd conditions, accuracy may be affected by crowding and occlusion, camera angle and distance, weather, lighting conditions, staff activity, and people who remain stationary or stop to take photos. As a result, the estimates may differ from officially reported counts and should be interpreted as indicators of pedestrian activity and movement patterns rather than exact or comprehensive pedestrian counts.
Traffic insights are based in part on third-party data and are subject to the provider's coverage, methodology, classifications, and data availability. They may not correspond directly with official agency-reported traffic conditions. All findings are provided for research and informational purposes and may be refined as additional data and validation become available. For more information about the methodology, please see here.
Select either cross-match pedestrian movements or the before-and-after pedestrian plaza comparison.
Edge-AI observations show how spectators move through three connected mobility zones. Start with the cross-match picture, compare matchday timing, then select an individual match for detailed ingress, egress, and estimated-count charts.
Three fixed camera observation areas capture complementary parts of the matchday mobility system.
Compare how arrivals build before kickoff and how departures unfold after the match.
Note: Estimated pedestrian counts are derived from computer-vision detections within fixed limited camera fields of view. Accuracy may vary with crowd density, occlusion, lighting, and weather. These estimates should be interpreted as indicators of pedestrian movement volume and patterns, rather than exact or comprehensive pedestrian counts.
Showing all matchdays with available pedestrian observations.
Each line represents one matchday. Estimated ingress and egress are aggregated into 15-minute intervals for a direct comparison of timing and volume.
Estimated pedestrian ingress in 15-minute intervals, aligned to kickoff (0 minutes).
Estimated pedestrian egress in 15-minute intervals, aligned to the match-end reference (0 minutes).
July 5 partial data: NJ TRANSIT Rail Area counts for July 5 represent partial data. Observations from 12:00–1:00 PM and the final 30 minutes of egress (8:30–9:00 PM) were unavailable, so counts could not be extracted for those periods.
Each line represents one matchday. Estimated ingress and egress are aggregated into 15-minute intervals for a direct comparison of timing and volume.
Estimated pedestrian ingress in 15-minute intervals, aligned to kickoff (0 minutes).
Estimated pedestrian egress in 15-minute intervals, aligned to the match-end reference (0 minutes).
Note: The June 27 pedestrian bridge count may be affected by reduced detection accuracy due to shadows and reflections from the wet roadway. The figure should therefore be interpreted with some caution.
Each line represents one matchday. Passenger alightings are aligned to kickoff and passenger boardings are aligned to the match-end reference so timing and rider volumes can be compared directly.
Estimated NJ TRANSIT passenger alightings in 15-minute intervals, aligned to kickoff (0 minutes).
Estimated NJ TRANSIT passenger boardings in 15-minute intervals, aligned to the match-end reference (0 minutes).
Note: Passenger counts are estimates derived from edge-AI cameras via automatic technology at the NJ Transit shuttle bus terminal entrance. Count accuracy may vary due to weather conditions, staff movement, and individuals entering the terminal solely to seek shelter. The data reflects NJ Transit terminal activity only and does not capture or represent the separate FIFA shuttle bus operations or passengers.
Combined NJ TRANSIT passenger movements and Pedestrian Bridge activity
Comparison note: Combined NJ TRANSIT totals use all-day Rail Area counts for June 13, June 16, June 27, and June 30; the July 19 final uses phase-aligned Rail Area totals. Pedestrian Bridge totals for June 13, June 16, and June 30 are phase-aligned, June 27 retains its current rain-affected estimate and caution note, and the final retains its corrected all-movement total because its ceremony and multidirectional crowd activity differed from the non-final matches.
NJ TRANSIT Rail Area + NJ TRANSIT shuttle-bus terminal
Combined total: NJ TRANSIT Rail Area estimated count + NJ TRANSIT shuttle-bus passenger estimated count.
Note: Computer vision-based estimates may differ from official NJ TRANSIT ridership figures due to staff activity, pedestrian movements near counting lines, limited camera coverage, and detection accuracy. They indicate movement patterns rather than complete passenger totals. For official ridership information, see the NJ TRANSIT Interagency Transportation After-Action Report.
Note: The June 27 pedestrian bridge count may be affected by reduced detection accuracy due to shadows and reflections from the wet roadway. The figure should therefore be interpreted with some caution.
Source: NYU C2SMART
This analysis compares a standardized occupancy-based congestion index that serves as a proxy for how crowded the pedestrian plaza became during selected events before and after construction of the new pedestrian bridge. The August 2, 2025, WWE event was selected as the pre-new-bridge reference because its high pedestrian-plaza congestion provides a meaningful benchmark for comparison. Three non-final FIFA World Cup matches serve as the primary comparison cases, while the Final is analyzed separately because it exhibited a distinct post-match pedestrian-flow pattern.
Completed in mid-2026 just ahead of the FIFA World Cup, a new pedestrian bridge was built adjacent to the existing structure over Route 120, connecting MetLife Stadium directly to the American Dream complex and its parking garages. This additional crossing was designed to increase pedestrian capacity and improve the flow of spectators before and after events.
Stoppage-time goals kept spectators engaged until the final whistle, resulting in a delayed post-match pedestrian surge.
Comfortable leads well before the end of the match coincided with earlier departures. Across the three observation areas, approximately 10% of observed egress movements occurred during the final 30 minutes before match end, shifting a portion of demand ahead of the main post-match surge.
Pedestrian Plaza congestion scores use a standardized index based on a common observed area across camera views. The index ranges from 0 to 100, where 100 represents the highest congestion level recorded at the venue in 2025-2026. The pre-new bridge reference peak was 84.2 during WWE SummerSlam Night 1.
Following construction of the new pedestrian bridge, peak index values were 66.3 for Panama vs. England, 71.1 for France vs. Sweden, and 60.4 for France vs. Senegal.
Despite attracting 80,000+ spectators, these peaks were approximately 16–28% lower than the reference event. Queues also remained continuously moving during the three matches.
The July 19 Final's peak index of 80.2 occurred approximately 75 minutes after match end and reflected a distinct combination of delayed departures, ceremony timing, pedestrian contra-flow, and stationary crowd gatherings. Its peak was higher than the other three World Cup matches but remained 5% lower than the pre-new-bridge reference event and 20% below the highest level observed across all monitored events in 2025–2026.
The shaded reference area shows the WWE congestion profile. Select legend items to show or hide the three non-final match curves, then move across the chart to compare standardized index values at the same minute.
The August 2, 2025 WWE event is retained as the selected pre-new-bridge reference for this comparison. The standardized index is calibrated across the broader 2025-2026 venue dataset, where 100 denotes the highest recorded congestion level.
The final is presented separately because the presentation ceremony, trophy lift, and atypical multidirectional crowd activity produced a different post-match pattern.
Hover or focus on any bar for its exact value and comparison with the reference before the new bridge.
Interpretation: Lower values indicate less intense or shorter-lasting congestion within the shared comparison window.
Meet the researchers and technology partners behind the World Cup mobility analysis.
This research project was developed by the New York University C2SMART Center in partnership with the New Jersey Sports and Exposition Authority (NJSEA) and AIWaysion, with support from MetLife Stadium. It builds on research initially funded by the SEMPACT Center through the U.S. Department of Transportation's University Transportation Centers (UTC) Program.
Director & Professor
NYU C2SMART Center, New York University
Assistant Director of Research
NYU C2SMART Center, New York University
Senior Research Associate
NYU C2SMART Center, New York University
Ph.D. Student
NYU C2SMART Center, New York University
Ph.D. Candidate
NYU C2SMART Center, New York University
Ph.D. Candidate
NYU C2SMART Center, New York University
Research Assistant
NYU C2SMART Center, New York University
Co-Founder & CEO
AIWaysion
Software Development Manager
AIWaysion
Director, Product Engineer
AIWaysion
Software Development Engineer
AIWaysion
Assistant Professor, Rochester Institute of Technology
Former Postdoctoral Assistant, NYU C2SMART Center
Machine Learning Engineer, StoneX
Former Research Scientist, NYU C2SMART Center
For any inquiries or feedback, please contact Dr. Jingqin (Jannie) Gao jingqin.gao@nyu.edu.
Pedestrian charts, counts, insights, and methodology notes will be added for this match when the data is provided.
This match is omitted from shuttle-bus comparisons because the required source data are not available.
Pedestrian charts, counts, insights, and methodology notes will be added for this match when the data is provided.
This match is omitted from shuttle-bus comparisons because the required source data are not available.
This match is omitted from shuttle-bus comparisons because the required source data are not available.