Skip to content
Traffica

For cities & DOTs

See how your streets and signals perform, from the vehicles already on them

Traffica turns the V2X messages your roadside units already hear into segment speeds, safety hot spots and signal performance. No new detectors to install, and no personal data in the results.

What you get

Network-wide answers from probe vehicles

Every product is filtered to your entitled area and built only from data your agreements allow.

  • Segment speeds

    Mean, median and 85th-percentile speed per road segment and direction, every 15 minutes, with sample counts and the posted limit.

    • GeoJSON for maps, CSV for analysts
    • Late data recomputes the same answer
  • Hard-braking hot spots

    Events flagged by vehicles (hard braking, ABS, traction and stability control) plus braking detected from motion at 0.4 g or more.

    • Events per segment and per km
    • Reported and derived kept apart
  • Intersection performance

    Signal performance measures in the style of ATSPM, computed from probe vehicles and the signal’s own broadcasts.

    • Arrivals on green, platoon ratio, stops, delay
    • Red-light-running candidates as aggregate counts
  • Signal data health

    SPaT and MAP broadcast rates, gaps, unknown timing and signal groups missing from the map, per 15 minutes.

    • Spot silent or misconfigured units
    • Outages show as lag, never as false metrics
  • Lane-level intersection maps

    Every MAP revision is kept: lanes, stop lines, allowed maneuvers and signal groups, as valid at any point in time.

    • Geometry changes become new versions
    • Vehicles matched to the map of their moment
  • Next

    Road weather

    Wiper state, air temperature and pressure, and traction events from vehicles, turned into road-condition products.

Segment speeds

A 15-minute picture of every segment

Each position is matched to the nearest road whose direction fits the vehicle’s heading, which picks the right side of a two-way street and the right road at an intersection.

  • 99.9% of test observations matched to a segment
  • 0.44 m/s mean error against ground truth
  • Cells under five vehicles stored, never published

Midtown segment speeds

Median speed per segment · 15-minute cells · 8:30 am–8:45 am

Times SqHerald SqBryant ParkGrand Central

8:30 am: 45 segments stop-and-go, 98 slow, 86 moderate, 1 free flow, 0 not published. Slowest: W 50th St, 8th Ave – 7th Ave, 2 mph.

8:30 am
  • Free flow≥ 70% of limit
  • Moderate45–70%
  • Slow25–45%
  • Stop-and-go< 25%
  • Under 5 vehicles · not published
  • Hard-braking hot spot

Intersection performance

Signal performance without loop detectors

Traffica places a virtual detector 100 m upstream of each stop line, works out when every probe vehicle would have arrived unimpeded, and checks that moment against the signal’s own phase and timing broadcast.

Arrivals on green
Within 4.85 percentage points of ground truth in benchmark tests.
Approach delay
0.71 s mean error against ground truth.
Red-light-running candidates
83% recall and precision. Always aggregate counts, never evidence about a vehicle or driver.

Intersection performance

Signal metrics from probe vehicles and the signal’s own broadcasts.

Arrivals on green
35%
share of arrivals with a known signal state
Platoon ratio
0.70
> 1 means progression is working
Stop rate
73%
vehicles that stopped on approach
Mean delay
35.6 s
vs. arriving unimpeded
Red-light-running
5
candidates · aggregate count only
By approach · PM peak (4–7 pm)
ApproachPassagesArrivals on greenStop rateDelay mean / p85
6th Ave northbound320
39%
69%34 s / 68 s
Broadway southbound331
40%
68%33 s / 67 s
W 34th St eastbound143
26%
81%40 s / 79 s
W 34th St westbound140
25%
82%40 s / 80 s
Arrivals on green through the day
0%25%50%75%100%6 am9 amnoon3 pm6 pm9 pm

Signal data health · 15-minute cells

4 cells not published · coverage under 0.9 · max gap 900 s
6 amnoon6 pm10 pm

A signal-data outage is lag, never loss: cells without enough coverage are held back, then completed when the delayed data arrives. Red-light-running candidates are aggregate counts, never evidence about a vehicle or driver.

Your data stays yours

You decide who else may use your roadside data

RSU data belongs to the agency that runs the units. In Traffica, every use of it is governed by a data agreement you control.

  • Purpose-bound

    Each agreement lists the purposes it allows, such as traffic management or road-safety research.

  • Resale is your call

    Allow or forbid resale, for example to insurers. The platform enforces it on every query.

  • Region and time

    Agreements and customer entitlements are bounded by geography and validity period.

  • Your own operations

    Run your own ODE if you prefer. Traffica supports several upstream connections per agency.

Getting started

From roadside units to results in four steps

  1. 1

    Connect

    Your roadside units, V2X Hub or your own USDOT ODE forward to Traffica over a secure tunnel. Each site is registered with its location and intersections.

  2. 2

    Map

    We import the road network and pick up each intersection’s MAP broadcasts, versioned automatically as geometry changes.

  3. 3

    Measure

    Speeds, hot spots and signal metrics appear within minutes, and are recomputed whenever late data arrives.

  4. 4

    Review

    Your team gets the dashboard and API for your area. We compare results with your counts and detectors together.

Bring Traffica to your corridor

Tell us about your corridor and your roadside units, and we’ll scope your deployment together.