Plot-X relational database implementation

As part of the DAWE Data Exchange Project, a relational database implementation of Plot-X is required. The exact deliverable phase is yet to be determined as there was no clear indication of which deliverable phase this falls under in the Data Exchange Standard contract with the DAWE.

Background design

Plot-X is largely an O&M implementation to support the capture and exchange of plot survey data. There have been many systems before it that successfully achieved what Plot-X strives to do, but at a smaller scale.

Veg-X

Veg-X, an XML-based data exchange standard comes very close to the requirements of Plot-X. Veg-X identifies the notion of making observations on features of interest at different granularity levels. It works well for its small scope (using AggregateOrganismObservation and ObservationGrouping), but the downside is that these aggregated feature entity types are fixed in the exchange schema and they have the potential in blowing up the schema entities as more ecological domains (such as soil, landform, land surface, etc.) are observed in the system.

ODM2

ODM2, a data model for earth observations, also based on O&M improves upon Veg-X by creating a features of interest entity as a controlled list rather than as separate entities for each feature of interest in the data model. This allows the data model to scale up for any number of features in the system without having the need to be extended.

ODM2 also has the notion of related features through its RelatedFeatures entity. This is particularly useful in representing parent-child relationships between features of interest in an ecological context. An example of this would be to relate the feature of interest stratum at a particular site to its plant population for the purposes of making an observation of the structural summary of the site and its most dominant species at the upper stratum.

The potential downside with modelling the feature of interest as an entity is due to the employment of an entity-attribute-value table for describing attributes of instances of the feature of interest, as it does not allow the enforcement of data types as well as a way to state if the attribute is required or not.

The main article describing features of interest in ODM2 can be found at https://github.com/ODM2/ODM2/blob/master/doc/ODM2Docs/ext_samplingfeatures.md

Implementation

Data models

A set of conceptual, logical and physical data models will be delivered for a relational database implementation. A reference implementation in PostgreSQL will be delivered.

Database tools

The implementation of the relational database model must support the the Plot-X data exchange format in terms of inbound and outbound transfer of data. To facilitate this interoperability, a set of reference implementation tools written in Python will be delivered.

Import

A tool to import Plot-X data into the relational database system.

Export

A tool to provide a database export of the data into the Plot-X data exchange format.

REST API

A REST API to provide a pull architecture for the Plot-X data exchange format. This may not be a required deliverable. The contract with DAWE states an API, but after reading it again, it is most likely for the mobile application and/or the controlled vocabularies (lookup tables).

Questions

There will be two databases - one developed by TERN Surveillance and one developed by TERN DSA (for Plot-X)

Yes, no?

It could be possible that the relational database system is the one which TERN Surveillance is already developing and we just need to provide the import and export tools to interoperate it with the Plot-X data exchange standard.

Clarify on the usage of controlled vocabularies and if they will be implemented as SKOS controlled vocabularies or as lookup tables in the relational database

We need to give examples of both as it is up to the adopter how they want to implement the controlled vocabularies.