Just-in-time synthetic computations on time-series data using a metrics approach

1 . A computerized method for just-in-time synthetic computations on time-series data using a metrics approach, comprising:

receiving a set of raw time-series data from a plurality of edge-device signals;

synchronizing the raw time-series data;

implementing a metrics feature that enables the creation of a plurality of synthetic metrics that operate on time series data, wherein the metrics feature analyzes and reasons about units of the raw time-series data and maintains correct unit types through mathematical transformations of the raw time-series data, wherein the synthetic metrics are values that result from combining other metrics, and wherein the metrics feature analyzes and reasons about units of the raw time-series data and automatically generates a correct unit type at the output of any computation;

enabling a query of the computed data just as a query of the raw time series data is performed, wherein the query results are visualized alongside the raw time series data in user-end displays segregated by appropriate units;

computing a plurality of synthetic metrics on the raw time series data, and wherein the step of computing a plurality of synthetic metrics on the raw time series data further comprises:

using a scalar value to convert the raw time series data to a set of metrics values;

providing a set of query results are visualized alongside the time series data in a dashboard view;

maintaining the computations in a synchronized state with the raw time series data is a complex task; and

determining the correct unit type from a combinations of units of the sensor data, and

wherein one or more of unit types of the sensor data are collected, and

wherein a multiplication operation is used with the scalar value to convert the raw time series data to a set of metrics values.

2 . The computerized method of claim 1 , wherein the raw time series data is obtained from a plurality of sensors.

3 . The computerized method of claim 2 , where the query of the computed data comprises a specific input pattern.

4 . The computerized method of claim 3 , wherein the data is segregated by appropriate units.

5 . The computerized method of claim 4 , wherein the correct unit type is used to provide a context and to a user-end displays of the time-series data.

6 . The computerized method of claim 5 , wherein the computing a synthetic data is dynamic and performed when the sensor data is obtained and batching on the backend is avoided.