Closed ElodieFZ closed 2 years ago
Make sure we document where the data comes from if we end up using an aggregated time-series, using provenance or history.
We request the default time-series only (so only one dataset per parameter/station) requesting these endpoints:
Until now, we have worked with all the raw data in ODA (kvkafka label). So for example, for Blindern, we have four different datasets for the air_temperature at 2m, as there are four temperature sensors at the station.
We should only publish the default time-series when several time-series are available for the same parameters. It should be possible to discriminate between these timeseries when the new Frost (v1) is available.