GINA's Amaru sensor data system ingests data into a MongoDB database from CSV or other formatted data files. Once in the system, the data can be processed to filter or correct the data and then this processed data is placed back into the database. This processed data can then be used for graphing, download, or served out through REST interfaces.
Amaru is now being used by the Sea Ice Group at UAF for the Barrow Sea Ice Mass Balance Site sensor data collection. Every 15 minutes data is collected from sensors placed on the ice northeast of Barrow, Alaska. This data is moved into GINA's system where it is archived and then ingested into Amaru. Amaru then processes the data, creates graphs of the data, and then makes the data and graphs available to external web sites for people to use.
The following graphs are being updated every hour as the data from Barrow is ingested and processed. Descriptions of the graphs can be found on the Barrow Sea Ice Mass Balance Site web page.
The system can be divided into several major components:
The ingestion scripts are triggered either from the command line or automatically from the REST API. All imports run asynchronously.
Processing scripts take the raw sensor data and then apply a processing chain to it, the results of which get ingested back into the database. Several processing tools are available, for instance, mean/median filters, and data adjustment tools. The tools can use the R Project for Statistical Computing package or ruby to process the data.
The scripts are ran asynchronously with in, or can be ran externally from the web interface to do automated processing of incoming data.
The alert scripts check the incoming data for predefined criteria and then alert the systems users if that criteria is matched. Very similar to the processing chain, an alert chain can also be attached to a sensor data stream. This alert chain is made up of one to many tools that check the incoming data. The system's user can choose how the alert is handled once a match is made, by either sending email(s) or by generating an alert in the web interface, or both.
Alerts can check for a sensor that is no longer communicating or can check for a particular event.
This interface is used by the data provider to manage the sensor platform and any processing done on that data. This Web tool provides interfaces to all of the major components of the GINA sensor system. It also allows the user to view the raw and processed sensor data so they may be compared and adjusted. Graphs of the data can be constructed with the built in graphing language. These graphs can then be served out via the public web interface.
This is the public face of the sensor system. This face allows web sites and external users access to the sensor data and graphs through various interfaces (REST API's).
These scripts generate graphs of the sensor data and are controlled through predefined templates. Currently the system supports line, depth, and profile graphs.
This is a catch-all category of scripts and tools that build products or give access to the sensor data.