
Multi-year historical trends and local model calibration to fine-tune accuracy for your region.
Physical counters deliver trusted data, but only where they're deployed, leaving most of the network uncounted.
Deploying, maintaining and retrieving physical counters requires staff time, equipment and repeat site visits, multiplied across every location you need to cover.
Even well-funded count programs typically cover only a fraction of the road network, often leaving other priority roads unmonitored.
Manual collection programs operate on seasonal or annual cycles. By the time data is processed and available, months may have passed and conditions may have changed.
SMATS iNode starts with a sample of anonymized trips from probe data, then uses machine learning algorithms to expand that sample into full traffic volume estimates, calibrated against your ground-truth counts.
Volume estimates across interstates, arterials, collectors and local roads, covering your full network from a single platform.
Annual, monthly, hourly and 15-minute volumes for both mid-block segments and intersections, across multiple years.
No seasonal collection cycles, data processing or manual data entry. Volume estimates are available on-demand through the platform.
Volume data feeds directly into all iNode modules, including risk ranking, congestion scoring and corridor analysis, from a single source of truth.
Probe-data volume models start with national estimates that work well in aggregate. Local calibration goes further, using your agency's own ground-truth counts to retrain the model so accuracy reflects the traffic on your roads.
Counts from your permanent and temporary count stations become the training data that tunes the model to your region. Use the built-in self-serve calibration tool or have SMATS do it as a managed service.
The model learns corridor-specific traffic behaviour, seasonal variation and regional characteristics that a national model averages away.
Each new round of count data refines the model further. Calibration is not a one-time event, it's a continuous process.
Manual counts capture what's happening on the ground. The calibration tool extends that local accuracy across your entire network, delivering scalable, defensible volume estimates.
Every new round of count collection refines the model. Accuracy becomes iterative and adaptive, always moving forward, never locked to an aging baseline.
Model quality is driven by your own data. You define what "accurate" looks like for your region, your corridors and your planning needs, with no vendor dependency.
Manual model calibration is slow and often impractical. The built-in self-serve calibration tool automates the process end to end. Update and improve models in minutes, not months.
| Capability | Traditional Counts | Standard Big-Data Providers | SMATS iNode |
|---|---|---|---|
| Network-wide volume coverage | |||
| AADT and multi-year trend data | |||
| All road types and functional classes | |||
| No physical hardware deployment required | |||
| Local calibration, Self-serve or as a managed service | |||
| Continuous model refinement as new counts are collected | |||
| Full control over model accuracy and timing | |||
| Calibration applied across all platform modules |
Maricopa Association of Governments, Phoenix, Ariz.
MAG, the MPO for the greater Phoenix region covering five million residents, needed to validate whether Floating Car Data could reliably support regionwide volume estimation.
SMATS built a locally calibrated model and evaluated it across 219 locations. The result: R² of 0.98, mean error of -1.5% and strong alignment across all road types and time periods.
.jpg)




















See how iNode delivers network-wide volume data and how iCalibrate lets you calibrate it with your own counts for local accuracy.
Book a Demo