This portal is a proof of concept built for EU AGRI. It shows how a self-service data-science platform (Onyxia) and a modern BI tool (Apache Superset) can be combined on a managed Kubernetes cluster to publish trustworthy agricultural market indicators — and how those dashboards can be securely embedded into a public-facing website like this one.
Architecture
Data scientists work self-service in Onyxia (Jupyter/RStudio), read and write datasets in MinIO S3, and publish dashboards in Superset. A scheduled Argo Workflows job can refresh the data automatically. This portal embeds the Superset dashboards through the official Embedded SDK; a small backend mints short-lived guest tokens so visitors need no account.
What it demonstrates
- Self-service analytics on a shared, governed Kubernetes platform.
- Single sign-on for analysts via Keycloak (OpenID Connect).
- Live market dashboards (pig, beef, dairy) sourced from the EU agri-food API.
- Secure public embedding: scoped, short-lived guest tokens — no anonymous access to the BI tool.
The platform
Onyxia
Self-service datalab (INSEE) — Jupyter, RStudio, on-demand services.
Apache Superset
Open-source BI: SQL exploration, charts and dashboards.
Keycloak
Identity & SSO (OpenID Connect) for analysts and the BI tool.
MinIO
S3-compatible object storage for datasets and artefacts.
Argo Workflows
Scheduled pipelines to refresh the market data.
This portal
Lightweight site embedding Superset dashboards via guest tokens.
Note
This is an unofficial proof of concept, not a service of the European Commission. Market data originates from the EU agri-food data portal (agridata.ec.europa.eu).