Easily query graphs and time-series data

Chrontext allows you to use a knowledge graph to query time series data in your existing database. This means you can use your existing models, expressed as a knowledge graph, to make time-series data accessible to your analysts. To make it easier to write queries accessing time-series, we have decided to add some syntactic sugar to the mix. Adding syntactic sugar means providing a convenient and simplified way to write something, avoiding repetition and improving accessibility.

In the query below, we are navigating a Solar Photovoltaic Power plant called "Eureka" to find every inverter, which converts DC power from the panels into AC power. Then, we find the time-series associated with the input DC power to each of these inverters, and specify aggregations every ten minutes. We have constructed an example Solar PV power plant using modelling principles using ISO/IEC 81346 (Reference Designation System) drafted by Statkraft, and a high resolution data-set from a NIST plant in Maryland, USA.

In the background, we are fetching data from a graph database, and from an existing database where time-series are stored (here: Google Cloud BigQuery). The query results are provided to us as a DataFrame, in a nice and tidy format.

Of course, there is no power production in the middle of the night, but since chrontext provides a DataFrame as output, this time-series can be immediately visualized.

We view the query above as a virtual data product, which can be shared with the rest of the organization. In a forthcoming post, we will look at how we can construct a simple materialized data product containing some derived time series data, which is again accessible with a query.