Phenodata and Grafana

[amotl]

About

At [RFC] New Grafana data source "BIENENJAHR", we are looking into how to optimally bring phenology data into our Grafana instance at https://swarm.hiveeyes.org/.

Details

We worked on bringing phenodata and Grafana together the other day, see README ¹ and the spot where phenodata is being used ².

The demo program focuses on flowering events for phenology species suitable as fruits for bees (Apis mellifera), and generates corresponding Grafana annotations. The location is Müncheberg, Brandenburg, Germany.

Originally posted by @amotl in #3 (comment)

Footnotes

1. https://github.com/panodata/grafana-pandas-datasource/tree/0.3.0/examples/phenodata-mellifera ↩

2. https://github.com/panodata/grafana-pandas-datasource/blob/0.3.0/examples/phenodata-mellifera/demo.py#L51-L92 ↩

[amotl]

With the resources shared above, we explored how to standardize the result channel from Python to Grafana by using pandas DataFrames as a canonical transport. However, this currently does not explore a way to standardize the input channel from Grafana to Python, i.e. the query expression, or a set of query parameters.

Now, with the Filtering with SQL feature, based on DuckDB, we would be on a level which would make any further thoughts about how to marshal the query expression or parameters from Grafana to Python obsolete, because it would be standardized on SQL, which is excellent.

In this way, we could hook phenodata fairly straight-forward as an SQL data source into Grafana, in order to finally approach an attempt to actually implement [RFC] New Grafana data source "BIENENJAHR". Based on this architecture, I think pretty much everything becomes possible with regards to universally supplying the various places in Grafana (chooser/selector, data tables, graphs and any other types of panels) with data.