DataVic Access Policy draft review and Salsa’s nine recommendations

Introduction

The Victorian Government is updating its DataVic Access Policy, which was endorsed way back in 2012. A lot has changed in the world of open data in the past eight years, and Salsa welcomes the review and update of Victoria’s policy. We also commend the release of the draft version for public consultation. It is vital the public and industry, who are key stakeholders, are able to provide feedback on the proposed policy.

The initial findings from this process are due to be released soon, and we thought we would take this opportunity to share our thoughts and experiences with the existing policy, and our take on the proposed new policy.

The importance of open government and open data

Salsa’s mantra is helping governments become more open, more connected and more consolidated. We believe open government is crucial to a healthy functioning democracy, and open data is one of the key pillars of open government.

“Inherently anything that is ‘open’ is more trustworthy than anything which is ‘closed’,” says Paul Morriss, Director - Projects at Salsa. “If citizens are able to see open data examples, and are able to realise data underpinning the applications they interact with are from open data sources, managed by government custodians, then an increased level of trust is built.”

Releasing data under an open, permissive licence is not only critical for government transparency and trust, it also drives innovation and promotes efficiency – people are not collecting the same datasets or trying to solve the same problems. Furthermore, the public and industry are not constrained in how they use the data. (For example, we analyse and visualise open datasets as part of our Open Data Insights series; last October, we looked at Victorian school zones data, which was released as a spatial dataset for the first time in 2019.)

“That is the great promise of open data – it lowers the barrier for innovation,” says Alfred Deeb, Founding Director at Salsa. “But we are still in the infancy of the open data movement. There is definitely greater awareness, and momentum is building. I feel like we are on the cusp of greater adoption, which is very exciting.”

Why it's time to update Victoria’s open data policy

In 2015, the Victorian Auditor-General Office (VAGO) conducted a detailed review of the DataVic policy. It found a lack of a systematic and consistent approach to categorising, storing and managing data, meaning the public were not provided full and open access to data.

Since the 2015 VAGO review, the Victorian Data Sharing Act 2017 was introduced to improve access to public data. In April 2017 Salsa Digital took over the management of data.vic.gov.au and implemented significant improvements to the portal, simplifying the process of releasing and managing data, and harvesting data directly from certain departments.

But while these enhancements were significant, there remains considerable roadblocks to achieving the full potential of Victoria’s open data. The Department of Premier and Cabinet (DPC) review of the current policy highlighted several issues:

• Open data was often seen as a burden by departments and agencies who struggled to find the resource capacity to release data under the current guidelines

• The existing policy does not provide data custodians with enough support and clarity for releasing data, and custodians were uncertain about the risks related to confidentiality, security and other restrictions for release

• It is difficult to quantify the economic value of open data.

Alfred Deeb agrees with these findings: “No one in government thinks open data is a bad thing. However, my observation is the departments do not have the resources or capacity to achieve what the policy sets out.”

To address these and other issues, DPC made 10 recommendations – starting with creating a new open data policy to “reflect contemporary data practices and align with terminology used by other jurisdictions”.

The DPC recommendations are listed below:

1. Establish a clear vision for open data by endorsing a new open data policy

2. Define policy outcomes to capture the full range of benefits of open data

3. Explore options to measure and maximise the impact of open data, including through collaborations with open data users and other jurisdictions

4. Provide clear direction as to what data agencies are expected to release

5. Consult with OVIC to mitigate the risk that open data could be used to identify individuals

6. Publish open data under permissive licences, with public domain being the default

7. Develop a category of ‘specialised data services’ to allow agencies to recover costs of releasing data to sophisticated users, where the costs of doing so are significant

8. Work with agencies to prioritise the publication of datasets that further policy outcomes and build release into their data management practices

9. Engage the public on the proposed open sata policy

10. Ensure policy and guidelines are easy to follow, relevant, and up-to-date

Highlights of the draft open data policy

One of the key recommendations of the DPC report was to provide a clear vision for the state’s open data policy and goals, recognising a need to cultivate cultural change across government towards more proactive publication of data.

The report also recommends broadening how we measure the benefits of open data. The draft policy suggests the benefits of open data should be to:

• Increase government transparency and public trust

• Enable data-driven economic and social development in Victoria

• Enhance public engagement with issues that affect the lives of Victorians

• Support responsible data stewardship within the Victorian Government.

But perhaps the most significant change is to the underlying principle of which data should be released. The current policy aspires to make all data open by default (unless there are concerns due to confidentiality or security). While this goal is admirable, in practice most departments simply do not have the resources to release all their data. Recognising this, the draft policy suggests that departments should “proactively” release open data.

Salsa Digital’s two cents (and our nine recommendations)

Salsa stands for all things open – open source, open platforms, open data, open design, and open government. This is why we are encouraged and excited by the Victorian Government’s update of its open data policy.

The update to Victoria’s open data policy is timely. A number of other jurisdictions have either updated their policies, or have released their initial policy more recently. New South Wales updated its open data policy in 2016, and are “prototyping and testing policies and frameworks” that may complement their updated policy. Queensland has had an open data portal since 2012 and has consistently published an open data strategy, but its official policy statement was announced in September 2017. Tasmania’s open data policy was released in 2016, and the Northern Territory went live with its open data portal in 2019.

We have reviewed the draft Victorian open data policy and have made nine recommendations. We believe these recommendations will increase the adoption of open data across government, improve the quality and relevance of the data, and increase its usage by the public and industry.

Recommendation 1: Provide agencies with clear, well-defined criteria for identifying high value datasets

We would love to see a time when “open by default” is achievable for government agencies, but recognise opening up government data is resource intensive and we welcome the change to “proactively release data to achieve the policy outcomes”. In fact, we believe this change could be beneficial to the public if there is a well-defined process of identifying high-value and high-impact datasets and prioritising these for release.

Recommendation 2: Require all agencies to publish and maintain a definitive register of all available datasets

The register should include the corresponding release status of each dataset – that is, whether it is already published, or is scheduled for release or is not suitable for release for security or privacy reasons.

Recommendation 3: Provide a mechanism for the public to request data

If the public and industry are provided with a mechanism to request unreleased data, the agency will get feedback from the end users on which datasets to prioritise for release.

Recommendation 4: All agencies should have a dedicated open data ‘champion’ to promote and drive open data adoption within the agency

Removing the “open by default” requirement should reduce some of the burden on agencies, but without a dedicated open data ‘champion’ within each department, Salsa does not see greater adoption across government.

“It's great to have a whole-of-government open data policy, and the changes outlined in the draft are important. However, it would be really impactful if each agency has an appointed champion to promote and drive open data adoption,” says Alfred Deeb. “The government would ideally create the space and funding to have a dedicated resource within each department to push and promote and drive it.”

Recommendation 5: Create a central body that supports and provides resources for open data ‘champions’ and open data adopters

An internal champion working alone in an agency will face considerable challenges, and we recommend creating a central body (or council) to provide support, tools and resources to the open data champions. By having shared resources and solutions, individuals are not trying to solve the same problems in isolation; there is less redundancy and greater efficiency.

To further enhance this recommendation, a national forum should be created where representatives from each state and territory can come together to share ideas, provide support and exchange learnings.

Recommendation 6: Ensure all published datasets include meaningful, accurate and detailed metadata

As well as producing and releasing data, we need to ensure that it is also used widely for the full potential of open data to be realised. And having accurate and detailed metadata is crucial to allowing greater usage of open data; it means potential users can start interacting with data immediately, without having to contact data custodians to clarify aspects of the data, such as what fields mean, or what units they are in.

Recommendation 7: Ensure all datasets adhere to consistent standards

Nathan Perry, a senior developer at Salsa and one of our resident CKAN experts (an open source data portal platform), says that having consistency of data will be crucial in getting industry to utilise open data and drive innovation. Having consistent data will also make it easier for the general public to interrogate and use the data – whether that is via open APIs or otherwise.

“It is difficult to build tools and resources for the public unless the data is consistent, and adheres to a common, open standard,” says Perry. “A consistent format means we can apply standard visualisation and analysis tools, which means users can gather insights quickly, without having assumed technical knowledge.”

Recommendation 8: Provide resources that make datasets easier for the public to find and use

There are many aspects to this, including – but not limited to – greater public education around accessing and using open data, easier discoverability of datasets, better documentation (case studies, detailed metadata) and building strong open data communities.

“Ease of discovery is an important aspect of driving value from data. The most insightful data is not useful if it can't easily be discovered, shared and used,” observes Paul Morriss. “The promotion of open data needs more focus. The technical building blocks to share data and build applications seem to be in place, but the potential user-base of such data is currently underrepresented due to lack of awareness. The cycle of promotion, adoption and increased community will help build the user-base.”

Jim Tasevski, who has been Salsa’s Engagement Manager on many CKAN projects, says providing tools to easily analyse and visualise data will greatly help citizens engage with open data. Without such tools, there is a technical and knowledge barrier for people to use the data governments provide; firstly they will need access to data visualisation software, and also some technical knowledge in transforming the data into a format that can be visualised.

“If we have data that conforms to well-defined standards, we can provide tools for users to easily visualise the data,” says Tasevski. “By providing such tools, users don’t have to worry about transforming data and importing data into analysis software. They can do simple interrogation and analysis of the data within the data portal itself.”

While not an official ‘how-to’ resource, we released a “how to publish and open your datasets” guide back in 2018. It serves as a step-by-step guide for authenticated Victorian public servants to register their datasets, for public discoverability and access, via the official Victorian Government open data directory (formerly referred to as the asset information register (IAR)).

Recommendation 9: Engage and collaborate with industry to drive innovation through open data

An important catalyst for greater industry involvement is the adoption of open APIs, and the Victorian Government has been leading the way on this front.

“Industry needs programmatic access, and open APIs will allow that,” says Alfred Deeb. “If you allow programmatic access to your data through open APIs, you lower the barrier for co-production and you will drive innovation.”

Open data is the foundational layer that supports open APIs. By allowing programmatic access to the data, industry can start to build solutions and automated systems.

Final thoughts

Open data is a critical component of open government and an important cog in the machinery of democracy – it promotes greater transparency and trust in government. But the potential of open data is immense. The opening up of government data facilitates the development of solutions to common problems; it reduces duplication and redundancy, and helps avoid silos of knowledge.

Open data not only fosters innovation, but promotes more efficient innovation. That is why Salsa believes the update to Victoria’s current open data policy is important and timely. We also believe our recommendations will help government agencies and end users deliver on the promise of open data.