A

• Actual - An actual measurement taken by a sensor.

E

• Execution Time - Execution time is the time at which an instance of Semaphore is run.

G

• Generated Time - Generated time is in reference to input data. It is the date time at which that Actual or Prediction was created. For an Actual its most likely when the sensor recorded that data point, for a Prediction it is the time when that predictive model created the prediction.

I

• Input - Input in reference to Semaphore refers to data going into a model. Data, including measurements and predictions, collected through data ingestion and requested by DSPEC input specifications.

L

• Lead Time - Lead time is the span of time between a predictive model is run and the time the prediction is computed for. For example, a model that predicts the water level in 24 hours would have a lead time of 24 hours. If it is run at 2pm on Monday, its 24-hour lead time prediction would be for 2pm on Tuesday. It is assumed that when the model is run, it uses the latest data and predictions available at the time of the prediction.
• Location - Locations in the context of Semaphore would be the location for a given data series. For example some longitude or latitude or some code-name for a location such as HCP for Horace Caldwell Pier. Location is one of the 4 descriptors Semaphore uses to differentiate between data series.
• Location Mapping - A location mapping is simply a relationship between a unique identifier for a location that Semaphore uses and the unique identifier a data source uses. Ex. Semaphore might call Packery channel packChan and NOAA tides and currents identifies the location as 8775792. Ergo a mapping exists between packCHan and NOAA's identifier 8775792.

O

• Output - Output in reference to Semaphore refers to data coming out of a model that Semaphore itself computes/operationalizes.

P

• Prediction - A prediction is a data point that has not been sampled from a sensor, instead has been generated by some predictive model.

R

• Reference Time - A reference time, in context to Semaphore, is a time at which a model is expected to be run. If an AI model is trained on hourly data, then its predictions are more than likely on the hour or shortly after the hour. Thus when Semaphore runs this theoretical model all of its input data should be selected in reference to the stated reference time, and not the time at which Semaphore was executed.

S

• Semaphore Series Description - A Semaphore Series Description is a collection of key identifiers that can be used to identify unique Output Series within Semaphore.
• Series - A series is one or more related data points that span a time range or a spatial range. A series can be uniquely identified by its location, data source, series (series name), and optionally a datum. Temporal information simply selects ranges of data from a series and not differing series. Series might also be in reference to it being one of the 4 descriptors Semaphore uses to differentiate between data series, the descriptor being some unique name for the series. Series that consist of Actuals should start with a d and Series that consist of Predictions should start with a p.
• Series Description - A series Description is a collection of four Descriptors that describe a unique data series within semaphore. The four descriptors are the location the data is referencing, the data source the data is from, the series name of the data, and optionally a datum.
• Source - A source is referencing the data source in which a series is from such as NOAA's Tides and Currents API or TCOON. It is one of the four descriptors Semaphore uses to uniquely identify data series.

V

• Verified Time - Verified time is the time in which an actual measurement was sampled or the time in which a prediction is for. For instance, if a model predicts the water level at some location in 24 hours, if you run this model at 2pm Monday, the prediction's verified time is at 2pm Tuesday, 24 hours later. You can think of Verified Time, as the point in time you would have to go to to verify if the prediction was correct.