This week, I’m at the 128th Open Geospatial Consortium (OGC) Member Meeting in Delft, Netherlands taking place from 25th to 28th March 2024. The theme for the meeting is Geospatial Data for Building Information Models (GEO-BIM) for the Built Environment, with various plenary sessions discussing overarching topics alongside parallel sessions supporting the activities of the OGC Data and Standards Working Groups.
As this meeting is taking place in Europe, Tuesday morning’s plenary focused on Collaborative Solutions and Innovation across Europe, showcasing the broader aims and supportive actions of OGC. Interoperability is a key focus of the OGC as it allows the broader community to save costs, open up new possibilities, and reduce risks. However, it is the goal, and not the means. The FAIR Guiding Principles – Findable, Accessible, Interoperable and Reusable – underpin this goal. Reusability is a key focus, for example., reusing existing software, instead of building new software from scratch.
A key question for reusability is Evaluation. Can the software be reused? Can I reuse it? Is the software out of scope? I addressed this point in my presentation, discussing the Training Data Markup Language for Artificial Intelligence (TrainingDML-AI) to support the reuse of machine learning (ML) training datasets within the Earth Observation Exploitation Platform DWG session. If we reuse training data and resulting models, we need to know their envelope of operability. Pixalytics is part of the TrainingDML-AI Standard Working Group, aiming to develop standards for the information model of AI/ML training data, including governance, metadata, classification, and reuse. L last year, the group published its first (conceptual) part as an OGC international standard.
One key topic that has emerged from several of the parallel sessions is how to share & reuse vocabulary so we know the details of the data we access, which the OGC addresses in part through the OGC Rainbow – a source of information about things (“Concepts”) the OGC defines, or that communities ask the OGC to host on their behalf. One role it’s taking on is being the Registration Authority for the OGC/ISO data quality standard for geographic information (ISO 19157-1), supporting Part 3, which will be the Data Quality Measures registration.
Picture of Delft
While travelling to the meeting, I was reading an Off to Lunch email by Business Leader, which showcased a column by Ed Smith, and it was an interesting juxtaposition with what I am discussing here. Ed talks about the expectation that “Big Data” alone can drive insights, forgetting the importance of storytelling. To back this up, he gives a Spotify example of how infinite choice fatigues people and results in narrowing rather than expanding their choices, alongside how it is difficult for data to inspire people on its own. This story came back to me while listening to Tuesday’s innovation session, where the discussion part included Common European Data Spaces with intermediaries to sell the data because providing the data on its own isn’t sufficient to encourage uptake.
All businesses, whatever they sell, and particularly with the tech and geospatial sectors, need to remember that they need to sell their story, which is partly why Pixalytics writes a blog! Once again, I’ve met people at this meeting who read our blog, which is lovely to hear.
We’re halfway through the OGC meeting, and we had great discussions so far. It’s also been wonderful to catch up with old friends, and meet new ones, from all sectors. One of the joys of OGC is the breadth of organisations and people who are part of it, making the meetings so interesting. Looking forward to the rest of the week!
