Services

KurrawongAI aims to support all aspects of Knowledge Graphs, including designing data, building systems, maintaining them and providing KG and system training. If there is a KG service you need and you don’t see it here, please let us know!

These are the services offered by KurrawongAI.

• Categories of service
• Products supported
• Example SLAs
• Arranging an SLA

Categories

KurrawongAI offers services in the following categories:

• Data Governance - policies, reviews and training
• Vocabularies
• Knowledge Graph data modelling
• Knowledge Graph system design, establishment, branding & maintenance
• Knowledge Graph-related software tool support
• Training

Products

Within the categories listed above, KurrawongAI provides Service Level Agreement-backed maintenance services for the following proprietary and open-source systems:

Data Governance

KurrawongAI is familiar with data governance conventions, best practice and policy for a range of data domains. We have several staff that have work on data governance for large Commonwealth data holders for decades.

KurrawongAI is able to offer the following data governance services:

• governance review
  • does your current governance regime do what you are expecting it to do?
  • likely not if you are conducting a review! but what exactly is missing in your regime, when compared with best practice among your peers?
• governance policy design
  • having developed and implemented many data governance policies for a number of organisations over the years, we are able to assess your needs and design policies for you
• governance implementation
  • after you have policies defined, you’ll need social infrastructure to implement it
  • we can assist you in creating the committee structures, role definitions and role allocations needed to implement your policy

Vocabularies

KurrawongAI has lots of experience in creating Knowledge Graph (SKOS)-based vocabularies, also known as taxonomies, to classify data and objects with.

Many of our vocabulary system establishment projects see us creating vocabularies for clients and then training them to create their own and to govern them (see Data Governance above).

Case Study

Vocabulary system establishment for South Australia’s Department for Energy and Mining

This is an example of a common government vocabulary establishment project.

We took Excel spreadsheets of 6 vocabularies of dataset classification terms created by the Department and converted them to SKOS vocabularies.

We also then creates 5 other SKOS vocabularies ourselves: several based on external vocabularies, one or two from scratch.

We then created an extension of the VocPub Profile’s vocabulary metadata to publish these 11 vocabularies with extended derivation and governance information.

Finally we established a Prez instance, branded for the Department, that published their vocabularies in both human- and machine-readable form for use by their staff, industry and cataloguing systems.

Use is on-going.

KurrawongAI is able to offer the following vocabulary services:

• vocabulary training
  • how do I make a SKOS vocab?
  • what is SKOS and what can it do?
  • how do I reuse existing vocabularies, rather than reinventing the wheel?
• vocabulary creation
  • we can make vocabularies from just about any source information, such as spreadsheets, databases, even emailed lists of things!
• vocabulary management
  • as per Data Governance above, we can assist you with policy and social infrastructure to manage your vocabularies
• vocabulary system creation
  • we have been installing vocabulary system tooling for 5 years now and can customise a toolset for you
• vocabulary analysis
  • we can analyse your vocabularies and implement best practice design
  • evn if we don’t know about your domain, we know about vocabulary data and use patterns

Knowledge Graph data modelling

KurrawongAI is extremely experienced with the creation and implementation of Knowledge Graph, including OWL / Semantic Web data modelling.

Case Study

Address Model for ANZ

This is an example of an OWL modelling and implementation project.

KurrawongAI responded to a request to implement a new Addresses data model for the Queensland government that catered for future address data requirements, as specified in the

ICSM Addressing 2035 strategy

.

In 2022 we produced an OWL model of address information derived from the

Geocoded National Address file relational model

and an OWL version of the ISO 19160 international address model that KurrawongAI staff developed 2017 while at CSIRO.

We implemented example data for this model within a Knowledge Graph system (Fuseki) and also a relational system (Postgres) to demonstrate the model s competence.

The work was well received by the client - the

Queensland Department of Resources

as evidenced by follow-on Solutions Architecture work in 2023 that designs the systems needed to implement this model in production.

We offer the following Knowledge Graph data modelling services:

• domain-specific data modelling
  • we’ve modelled in many domains, such as archiving, environmental science, spatial data and provenance
  • we are able to pair up with your domain experts and develop models with them
• enterprise data modelling
  • so-called “Supermodel” modelling whereby we model a broad scenario’s requirements as a data model
  • Supermodels are compartmentalised allowing for staged implementation
• model skill demonstration
  • query-based competency testing of models
  • so you can demo the power of KG models
• example data generation
  • we can hand-make data to assist with model demonstration
• data transform
  • from existing systems to new KG models
  • see Knowledge Graph tooling below

Knowledge Graph systems

We implement all aspects of Knowledge Graph systems, from RDF generation scripts to databases, adaptors for external systems to work with RDF data and APIs to deliver Knowledge Graph data online as Linked Data.

The main tasks we generally undertake for KG systems are:

• establishment
  • specification of infrastructure
  • configuration, such as index creation
  • containerisation and/or cloud infra coding
• getting data in
  • converting non-KG data into RDF
  • data loading API configuration
• getting data out
  • making custom APIs with tools like Prez
  • see Knowledge Graph tooling below for what we can do to link your external systems with KG systems
• management
  • of large and/or growing data
  • of complex data
  • see Olis, our very own graph management API

We provide commercial support for several RDF databases (also known as triplestores), in particular Fuseki. See the list of supported products above for all options.

Case Study

Geofabric Linked Data System

This is an example of a large-scale, operational, Linked Data system.

KurrawongAI staff extracted hydrological catchment data from an offline spatial data source using RDFLib and created Knowledge Graph data (RDF) from it, based on the Geofabric Ontology we designed for it. We then loaded the data into a KG database ( Fuseki) and then delivered the information online using Prez.

This Geofabric Linked Data System is part of Geoscience Australia’s KG-based

FSDF Data Platform

which is being built to support their Digital Atlas of Australia

We continue to support the operations of the Geofabric Linked Data System and expect to enhance the data it delivers in mid-2023 due to international interest in the delivery of spatial Semantic Web data.

Knowledge Graph tooling

We have several staff who have been long-time developers of the open source RDFLib Python RDF manipulation toolkit, including Nicholas and Edmond who are RDFLib maintainers. We use it to convert non-KG data into RDF for loading into graph databases. We also know well and use the Jena framework’s tools for this.

In addition to these tools, we maintain an RDF conversion and validation tool based on RDFLib, imaginatively titled RDFTools.

Both RDFLib and Jena tools can be used as libraries within scripts and workflows to generate RDF data from non-RDF sources, such as relational databases, spatial data files and so on. We have many instances of custom data pipelines that we’ve developed to create and update RDF data using these tools.

We also maintain a series of RDFLib-derived tools, such as pyLODE and VocExcel.

We offer the following Knowledge Graph tooling services:

• creation and application of RDFLib and Jena-based tooling
  • if you need a custom script/program to create or manipulate KG data, we can probably make it using RDFLib or Jena
• custom product creation
  • if we find that a custom script/program is highly reusable, we could package that up for you as a product
  • this is what has happened with RDFtools: we made this and a customer of ours has their own in-house copy of it, tailored for their needs
• supported use of pySHACL, pyLODE RDFTools & VocExcel
  • we’ve developed these tools and are happy to assist you with their use

KG tooling deployments are usually one-off things. Please contact us for a quote for your needs!

Training

KurrawongAI provides training services for the following domains:

• vocabulary creation & management
  • see Vocabularies above
• data strategy
  • how to plan for your organisation’s data in the medium and long term
• data governance
  • how to manage your data
  • see Data Governance above
• KG data generation
  • how to make RDF data manually or from existing, non-RDF data sources
• system management
  • how to maintain Knowledge Graph systems like Prez APIs and RDF databases such as Fuseki

We offer the following training options:

Example SLAs

We will soon (Feb 2023) bt placing example Service Level Agreements here to give you an idea of specific SLAs.

Arranging an SLA

Please just contact us directly in the short term. Eventually we will provide a decision tree tool here to help you define some of your requirements up front.