As of July 2026, New York’s congestion pricing program continues to generate useful evidence for cities considering demand management policies. One recent finding stands out: a New York City Department of Transportation study reported that congestion pricing did not produce the feared surge in parking outside the Manhattan charging zone.
For traffic engineers, the larger lesson is not only about parking. It is about measurement. When a major policy changes travel behavior, agencies need to know whether slowdowns appeared, where they appeared, when they began, and whether they lasted for minutes, hours, weeks, or an entire season. Averages alone cannot answer those questions. Ticon’s traffic analytics methodology is built around that exact problem: converting continuous speed and volume observations into a time-resolved picture of network performance.
A corridor may look acceptable when evaluated by average daily traffic, yet fail repeatedly during a narrow Friday afternoon period. A parking policy may create no visible daily surge, yet still shift arrival patterns by 15 or 30 minutes. A signal retiming project may reduce delay in the morning while worsening it in the evening. These are not edge cases. They are the normal operating conditions of urban networks.
Ticon addresses this by analyzing traffic as a sequence of time states rather than as a single daily number. In the Ticon Traffic Intelligence Report framework, each GPS probe represents a 15-minute interval, typically yielding 96 probes per day. Congested hours are defined as periods when the median speed falls below 80% of the expected speed for the road segment. Rush hours are defined more narrowly: median speed below 50% of expected speed, combined with traffic flow above 60% of the segment’s maximal capacity. Congested and rush-hour ratios are then calculated as a percentage of normal travel hours, from 6:00 am to 11:00 pm.
That structure allows analysts to identify not just that congestion occurred, but its onset and persistence. If median speed crosses below the 80% threshold at 7:15 am and does not recover until 8:45 am, the slowdown has a measurable start, duration, and recovery point. If the same pattern repeats every Tuesday and Thursday but not on Monday, the issue is no longer a vague “peak period” problem. It becomes an operational signature.
Traditional traffic studies often miss the beginning and end of slowdowns because they observe only a small slice of time. Ticon’s methodology was developed to overcome that sampling limitation. According to the Ticon Methodology document, the platform provides almost complete road network coverage, more than 97% of roads with functional road class 6 and above, and 100% time coverage by integrating permanent detectors, portable counters, GPS data, connected vehicle data, GIS information, demographic inputs, traffic organization data, and event transformations.
After cross-verification, filtration, and proprietary processing, Ticon estimates speeds, volumes, and derived traffic metrics for 95% of roadways at high spatial and temporal resolution. The road segmentation can be as short as 35 feet, with an average of about 225 feet, and time intervals can reach 5 minutes, and in many cases 15 seconds. This matters because slowdowns often begin locally: at a signal approach, near an on-ramp, at a curbside conflict point, or at a short section affected by turning traffic.
The report Data Ampleness, Accuracy and Reliability: Site Selection and Location Traffic Analytics illustrates the cost of partial observation. In one example, a one-week portable count in December closely matched Ticon during that week, differing by only 1.4%. Yet the resulting AADT estimate differed from Ticon’s year-round estimate by 17.6%. Ticon practice has observed differences above 75% in some cases. The problem was not that the counter was wrong during the week it operated. The problem was that the week was not representative of the full year.
The same principle applies to slowdown duration. A two-day count can identify a queue if it happens during the observation window. It cannot reliably distinguish a recurring weekday bottleneck from a short-term disruption caused by weather, construction, special events, or school schedules.
To understand when a slowdown began, an analyst first needs a defensible baseline. Ticon defines normal speed as the average speed of a vehicle traveling at maximum safe speed while obeying the speed limit, road signs, and traffic signals. That baseline is then compared with observed median speed at fine time intervals.
This approach is important because a slowdown is not simply “traffic moving slowly.” A 25 mph median speed may be normal on a downtown arterial with signals and pedestrian activity, while the same speed may indicate a severe disturbance on a suburban expressway. Ticon’s method connects speed interpretation to the geometry and expected operating condition of each road segment.
The platform also combines speed with volume and saturation analysis. This distinction is critical. If speed drops while volume is low, the cause may be an incident, work zone, weather event, parking maneuver, or signal malfunction. If speed drops while traffic flow exceeds 60% of maximal capacity, the segment is closer to a rush-hour condition. In engineering terms, the same speed curve can imply different remedies depending on whether the road is undersaturated, saturated, or oversaturated.
Duration is where continuous coverage becomes essential. A slowdown that lasts 15 minutes may call for monitoring or a minor operational adjustment. A slowdown that repeats for two hours every weekday may justify signal timing changes, turn-lane evaluation, access management, or demand management. A slowdown that persists across months may indicate a structural capacity or land-use issue.
Ticon’s congestion analysis uses time bins to measure the persistence of below-threshold conditions. Instead of describing a corridor as “congested in the afternoon,” the analyst can identify the exact time window, compare it by day of week, and test whether the pattern is stable across months. The retrieved Ticon TrafficScope materials describe graphs built for each road section and each day of the week with up to 15-minute bin resolution. This is the level of detail needed to separate a recurring bottleneck from a one-time anomaly.
Ticon’s before/after analysis extends the same logic to policy and infrastructure evaluation. TrafficScope compares traffic flow conditions before and after a mobility measure, using high-resolution datasets with 100% temporal and spatial coverage. In the Atlanta Smart Corridor case described in TrafficScope Before-After Analysis as a Mobility Improvement Tool for Smart Cities, the public expectation was a 25% travel time reduction and 40% traffic delay reduction after adaptive signal control implementation. Ticon’s analysis found a more nuanced outcome: average daily gain was reported at 3.12%, AM peak gain at 4.47%, and PM peak gain at 0.46%. The matrix also identified degraded time slots, including a problematic Friday afternoon period from 4 pm to 6 pm.
That example shows why duration and timing matter. A project can look beneficial in aggregate while still creating localized or time-specific degradation. Without a matrix by hour and day, the Friday 4 pm to 6 pm problem could be hidden inside a weekly average.
Once the start and duration of a slowdown are known, the next question is where it forms. Ticon’s Street Cardiogram method converts delay values for each segment into an along-the-street performance graph. The method reports delay in seconds per vehicle, level of service, and whether a segment improved, degraded, or remained unchanged after an intervention.
This provides a practical bridge between analytics and engineering work. If delay begins at the same signal approach every weekday at 4:10 pm, then spreads upstream over the next 20 minutes, the likely intervention differs from a corridor-wide demand surge. The first case may require signal phase adjustment, coordination review, saturation analysis, or turning movement evaluation. The second may require broader network management or demand strategies.
Ticon’s mobility improvement research also shows why reducing demand alone may not solve delay. In Traffic congestion: what works, what doesn’t, Ticon examined traffic flows at 126 intersections in nine U.S. states using about 200 million datapoints during COVID-era traffic restrictions. The study observed traffic demand reductions of up to 30% and more, but delay reductions on signalized roads were much smaller. In some cases, delay barely changed even when demand fell by almost half. The implication is clear: the start and persistence of slowdowns often depend on traffic organization, not only volume.
The same document notes that, based on Ticon experience, travel delay reductions up to 50% are achievable through signal timing optimization alone. But that potential depends on knowing which time periods and locations are causing the loss. Optimization without time-resolved diagnosis risks improving the wrong hour or the wrong intersection.
Fine-grained analytics must still be accurate enough for planning and operations. Ticon’s field studies provide useful benchmarks. In Data Ampleness, Accuracy and Reliability, Ticon reports that AADT estimation error can be kept within 20% boundaries with 90% confidence. Median average percentage discrepancy between Ticon hourly volume estimates and direct detector measurements varies between 5% and 20%, depending on infrastructure data availability. For 15-minute traffic flow volumes, Ticon reported a median absolute percentage error of 11.24%, compared with 6.5% for a pre-calibrated video detector under the same circumstances.
For congestion timing, this level of temporal coverage is often more valuable than a perfectly measured but brief sample. A detector may provide precise observations at one point. Ticon’s approach provides a broader empirical context: whether the slowdown appeared before, whether it repeated afterward, and whether it propagated across nearby segments.
The NYC congestion pricing parking finding illustrates a broader challenge for every city considering pricing, curb reform, signal upgrades, bus priority, or intersection redesign. A policy may not create the feared system-wide effect, but it may still alter traffic in narrow windows or specific locations. Agencies need tools that detect those effects before they become public frustration or before a promising project is judged by incomplete evidence.
Understanding when slowdowns began and how long they lasted turns congestion from a complaint into an engineering object. It defines the baseline, the threshold crossing,the duration, the recovery, the affected segments, and the likely operational cause. That is the foundation for practical decisions: whether to retime a signal, adjust curb regulation, prioritize an intersection, evaluate saturation, or verify whether an implemented measure delivered the promised result.
Ticon’s contribution is the combination of continuous temporal coverage, high spatial resolution, multi-source cross-verification, speed-volume analysis, saturation evaluation, and before/after reporting. In a network where conditions can change within minutes, that combination gives planners and engineers the evidence needed to see not only where congestion exists, but when it starts, how long it lasts, and what can realistically be improved.
Join the network to access Ticon case studies and to receive Ticon newsletters
@2023 Ticon
