For the last three years running, Link Digital has been both a sponsor and an active participant in the Canadian Open Data Summit. It’s a great open data event, built on a great open data community. In addition to inspiring Link Digital to continue our work using open source technologies to build and secure data hubs and portals, this year’s CODS in late November provided a lot to think about, on everything from the state of open data to AI.

Organised by the Canadian Open Data Society, this year’s CODS was combined with GovMaker Canada, an initiative led by the Pond-Deshpande Centre at the University of New Brunswick, focused on advancing public sector innovation and open government.

For our Technical Delivery Manager Lex Du Plessis, the University of New Brunswick presentations were a Summit highlight.

They really gave the conference a bit of a different vibe and allowed me to listen in on conversations that gave a different perspective on how data is used and some of the struggles that are being experienced by the people that work with data. They really brought an academia perspective to some of the conversations around how data is generated and what you could do with it. I read STEM field papers and listen to podcasts covering these fields and having these guys present in the way that they did overlaps with my interests.

Artificial intelligence dominates discussion

As has been the case for most open data events Link Digital has taken part in this year, AI was the dominating topic.

A lot of the conversation at the conference was around how to use AI, using AI to help with data discovery, using AI for data insights and making data more accessible with AI by using natural language,” says Lex. “My feeling is everybody is working on this and looking to incorporate some form of AI into their ecosystems.

A second thing about AI was that there really is an uneasiness in these discussions. People are really scared of losing their jobs and I can completely understand their sentiment. Narrow AI is really seen as a tool and can support users as per point A, but General AI is seen as a bit of an intrusion and the fact that it’s being included into everything (Google, Microsoft, etc.) really makes people unconvertable.

Moderated by Link Digital Executive Director Steven De Costa, on ‘The Spooky Physics of Shared Understanding.’ It featured Lex, Wilfreda Edward of the Canadian Centre For Nonprofit Digital Resilience , and Philippe Fortin, a Data Governance Specialist at the Autorité régionale de transport métropolitain, discussing the  paradox of how observing something changes the thing observed, and how this relates to open data practice and the shaping of our social and civic understanding of truth.

Senior Solution Architect Ian Ward also touched on AI in his presentation on tools and best practices for publishing and working with data dictionaries, including some benefits of publishing machine-readable formats.  

I’m very happy with how the session went and was pulled aside at the conference afterwards a number of times to discuss data dictionaries. The technologies presented were new to many people attending, and collecting and sharing provenance, validation and other metadata along with published datasets was new to some.

Publishing important metadata as data dictionaries along with the data clarifies limitations, linkages and restrictions on that data. These insights let users make informed decisions about how the data is used and what conclusions that can be drawn from analyzing it. Machine readable data dictionary formats allow systems to align controlled vocabularies, implement compatible validation and republish key metadata about data they consume. Like CKAN, data dictionaries give users the context they need to understand, effectively integrate and confidently build solutions with data.

Threats and opportunities

The Summit underlined both opportunities and threats facing the open data movement going forward.

“Open data has reached a level of maturity where we’ve moved past convincing people, and we’re now focused on remaining issues and getting the real work done,” says Ian.

There’s always a danger of backsliding though, especially when the path of least resistance is working in secret and not releasing data when it might lead to criticism or embarrassment. Many institutions have established patterns of working in the open. We should continue to draw attention to successes that are only possible by working this way so that these patterns become examples for everyone to follow.

High quality data with rich metadata is just as useful for AI tools as it is for humans. Data validation and controlled vocabularies can keep generative AI results from colouring outside the lines and metadata allows building queries or generating summaries with AI tools without breaking the bank by training models on every row. As AI features are integrated into any and every process the demand for clean, well described data to power these tools is only going to grow.

To me, it seems like open data has made some progress and has also slipped at the same time,

contends Lex.

My takeaway from the conference is that people get the importance of open data. They are relying on it more and more, but this also seems like a two edged sword. I think the problem that might be starting to pop up is too much irrelevant data or data not really capturing the information that is really needed.

The example that comes to mind is the most amazing dataset created by the province of Nova Scotia. The Engage Nova Scotia asked residents over 200 questions about how they feel about their province and the data that came out of that is just mind blowing. They are gearing up next year to repeat the survey and my question really is, why are we not doing this on a federal level. The amount of informed policy and decision making that can be based on this data would be amazing.

Link Digital: committed to the Canadian market

Link Digital’s participation in this year’s CODS comes on the back of the gradual success of our Canadian subsidiary, Datashades Ltd. According to Steven:

We have invested in Canada in the same way we have grown into our natural investment within Australia’s open data ecosystem. We secured clients in Canada in 2020, and we wanted to make sure we were there with all the authenticity and local ambition that has marked our history here in Australia.

Lex agrees:

Link Digital really has the opportunity to position itself as a trusted partner when it comes to open data in Canada and globally as we navigate through this new world that is evolving around us.

Talk to us about your data management capabilities.Contact Link Digital to discover how our expertise can benefit your organisation.