Raw observations flow in. Five extraction signals refine them. A single trajectory score emerges — showing not where something is, but where it is headed.
Every trajectory is decomposed into five fundamental signals. Each captures a different dimension of change — together they paint a complete picture of momentum.
The direction and steepness of the overall trend. A best-fit line drawn through every observation, revealing the fundamental trajectory.
Drag the dots to reshape the trend line
How quickly the rate of change itself is changing. Rising acceleration means the climb is getting steeper; falling means it is leveling off.
Watch how the rate of change evolves
The spread and jitter in the data. Low volatility means predictable movement; high volatility means the path is uncertain and erratic.
Left: smooth and steady. Right: noisy and wild.
An exponentially weighted recent average that gives more importance to what just happened. Early warnings of shifts appear here first.
Green line smooths the raw data with recency bias
Sudden regime shifts where the trajectory fundamentally pivots. These breakpoints mark the moments that matter most.
Orange markers flag detected shift points
Five extracted signals feed into a weighted combination. Adjust the influence of each signal and watch the trajectory score — and its prediction cone — respond in real time.
The scoring engine is domain-agnostic. Plug in a module for any vertical — skills, commerce, healthcare, insurance, education — and the same trajectory math applies.
Trajectory scoring runs in real time on production-scale datasets. No batch processing, no waiting — scores update the moment new observations arrive.
This product is one of 22 specialized solutions in the DobePros suite—a sentient infrastructure that evolves through real-world execution. By leveraging a unified, self-learning DNA for data, security, and commerce, every tool benefits from cross-platform hardening and collective technical resilience.