One of the objectives of such an approach is to apply the concept of FAIR data, by ensuring that data from the Water Information System is Findable, Accessible, Interoperable and Reusable, so that all actors from the public and private spheres can benefit fully from the data. These notions of simplification and reuse were defined not only by the producer but also by users, through a systematic joint construction approach involving testing of beta versions and a GitHub forum to benefit from users' feedback.
The API developed in 2018 aims to meet the high demand for data from the HYDRO Central Platform (PHyC), operated by the Central Hydrometeorological and Flood Forecasting Support Service (SCHAPI), which stores near-real-time measurements from approximately 3,000 hydrometric stations (French measurement network, operated by the DREALs or other producers such as local authorities). When it is complete (2020), the platform will also include "delayed time" data to replace the HYDRO database.
The monitoring networks, whether for surface water or groundwater, have their own infrastructure enabling data to be made available in near-real time. In 2018, in partnership with SCHAPI, the hydrometric reference database and the observation data regarding water levels (H) and flow rates (Q) were made available via a programmed interface (API) to ensure accessibility and performance for all through the Hub'Eau platform.
A processing chain was designed to query a set of SCHAPI web services every two minutes. After internal tests at SCHAPI, it was proposed to launch the production of the public beta version with APIs for sites, stations and raw real-time observations in September 2018. The results of queries are displayed in XML SANDRE format in addition to the JSON and CSV formats, which was an urgent request from the producer, itself a consumer of the infrastructure for its own needs.
Using the interface
Based on an infrastructure and methods adapted to the processing and storage of big data, the Hub'Eau APIs guarantee optimum availability. The new API integrated in 2018 can provide “near-real time” water level observations (because they are updated by their producer every 5 to 60 minutes in the source platform, SCHAPI’s PHyC). Hub'Eau queries the data source every two minutes and keeps observations available for one month. This API became a great success as soon as it was made available. Indeed, there have been in the region of 2 million calls per month on average, from the first few months of opening, making this API the most widely consulted, and above all, the one that centralises nearly 90% of data calls.
- Central Hydrometeorological and Flood Forecasting Support Service (SCHAPI), General Directorate for Risk Prevention of the Ministry for the Ecological and Inclusive Transition