Data quality

[mdutoo]

(to be spinned off in several issues)

1 a priori data quality, using an evented business rules engine.

• OPT conf'ble behaviour when fails: HTTP error, kept but with odq:valid={failed, timestamp} Data quality - incomplete data (Model non-compliance properties) #21 , kept but only in contributions...
• either custom using ex. Javascript engine and conf + test web UI in DC Playground, OR / AND by integrating an existing Data Quality solution (ex. Talend) or separate components (ex. Google Open Refine, common address checking web service...)

2 a posteriori data quality, more like business KPIs on resources / aggregation of resource data quality at top (project) level (and integrated in Playground project portal, OPT with high level quality threshold alerts), or for models that don't have a priori DQ enabled