Roadmap and rationale

Presention and rationale

...Some "what-and-why" items, explaining project's moivation... Or jump to roadmap.

Transfer standard Datasets to a PostgreSQL database, transforming it into fast, reliable and compact data-representations (CSV lines as JSONb arrays).
Offer "standard SQL VIEWs" (standard table names and column names) for the dataset SQL-representation.
PS: so also offer a simple universal data-transformation language (SQL) to describe, in a standard and reproductible way, the data provenance of my datasets (and perhaps any standard dataset).
Do class modeling of all datasets. (illustred below from sql-unifier/src/Appendix.).
Build (easy to) new datasets, as a mixed-datasets, from standard ones, knowing or modeling relationships (SQL JOINS, etc.). See illustration bellow.
See standard Datasets and Datasets-BR as an "ecossystem of reliable data" that have long-term digital preservation at git repositories, and an enhanced access by standard class modeling, standard API interfaces (as GraphQL with PostGraphQL), or consuming it as standard SQL tables.
...

To be portable: all datasets in one table.
To be easy for me: I like SQL, and there are costs of test and learning to use other (non-standard) tools like CSVkit.
To be easy to produce new mix-datasets, in a realiable and standaerd way.
To be easy to use the Datasets in PostgreSQL databases, as an independent, refreshable and flexible SQL-SCHEMA.
To be easy to offer the Datasets for my users, in standard APIs.
...

... next steps? ...

Consensus about a "SQL-kernel" framework to do all basic things.
Test with all
Test with other people and import/expot/analyse tools as CSVkit or Goodtables.
Test with all datasets of http://github.com/datasets
API: consuming online the datasets by GraphQL's interface of PostGraphQL project... Or others like http://postgrest.com
Expand and finesh.