Systems Thinking And Promoting Data Integration For Better Decision Making
Featuring in this episode is Engo Simonis, the Chief Technology and Innovation Officer at the OGC (Open Geospatial Consortium). Originally an ecologist, he later transitioned to a computer scientist, and eventually a systems engineer. At the OGC, he is part of a society that champions systems thinking and data integration to promote collaboration for better decision-making. OGC explores the role of geospatial in different systems – be it natural systems, human systems, ecological systems, or any other system. Quoting Engo’s insights,
“Computer systems are just an underlying layer, but what it really is – a system of humans that use computers.”
Digital Twins in Geospatial
Digital twins are digital representations of the natural environment. Creating a virtual representation of the natural world enhances how we understand and communicate what is happening in the environment. From this understanding, we can make informed decisions on how to influence the natural world for the better good.
Modeling natural phenomena provide useful insights into decision-making processes. But it is impossible to model the whole Earth at once – the Earth is a very complex system that comprises many subsystems. Instead, certain elements within the Earth system can be observed at a time.
For instance, we can only monitor temperature at a couple of places but not everywhere. Separate individual systems are needed to observe different physical properties at the same time. Then by integrating these individual pieces into a master system, we can obtain a complete holistic view of the whole system that gives realistic representations of how systems behave in relation to one another; and how we can effectively influence a certain system.
Open Data Access and Integration Challenges
Promoting open data access will also go a long way in improving the decisions we make on how we can best influence reality. Allowing stakeholders to discover and use data from different sources within their applications can generate new knowledge about our environment and hence the better the decisions we can make.
However, integrating datasets from different sources has its own fair of challenges. Despite possibly coming in all sorts of different formats, the data may also vary temporally. If there are no standards to guide how data from different sources can be combined, then integration can be a very big challenge. Over the years, the OGC as a society has been a big player in developing industry standards that promote integration between different datasets.
How Do Digital Twins Help In Decision Making?
The most effective decision-making process must be sufficiently informed with the right data. As a convention, decision-makers relied on printouts such as tables and charts to inform the choices they make about cities, or on how to manage global concerns such as climate change. But relying on static printouts is not the most optimal way to inform important decisions. Digital twins, which can provide different perspectives of a phenomenon, have quickly become an important piece in decision-making processes.
In using digital twins, it is possible to view different scenarios of what will happen if a certain decision is made. Additionally, combining digital twins with other simulation models of human behavior and responses of the environment to different conditions further enriches our understanding of how impactful a decision would be. Creating visualizations of the natural world makes it easier to see what is in the environment and as well understand why things are how they are.
Data Providers Vs Consumers in Modeling the Physical World
For data providers, the big challenge is that the consumer side, which they seek to serve, is not homogenous. Consumers may have very different views of the same phenomenon depending on their specific needs.
A user of a power grid system who uses power to charge their home devices will have a completely different view of the power grid compared to another one who needs to maintain it. Understanding how to serve data to various groups of consumers in a way that is economically efficient is critical for providers.
For consumers, their data needs largely depend on whether the problem they are trying to solve is simpler or more complex. A simpler problem that relies on less variables requires less data and is easier to model and implement. But complex problems like climate change require very large data sets – which may not necessarily mean big data, but large in the sense of lots of different data, and possibly more integration challenges.
Can There Be One Perfect Model Of Reality?
Both natural and human-made environments are very complex. Modeling either of them always involves some simplifications and reduction of the input variables.
The biggest challenge is not picking the right variables, producing the perfect model, and interpreting them in the right way. Because there is no one perfect model. People may have varying perceptions of the same reality due to a difference in contextualization and past experiences. Producing a perfect model that suits all interpretations of reality is impossible. Rather, it is realistic to create a subset of models for the most likely realities of the upcoming future.
The Technology Vs the Cultural Stack
The technology stack – the technology that we use to create models of reality is very dynamic. Rapid changes are eminent in this space as new advancements are realized from time to time. On the other hand, humans have behaved the same way for centuries without any major changes. The way humans learn has not changed over the years. Hence, the cultural stack is a constant allowing experts to focus more on developing better technology.
Humans and machines interact differently among themselves. Unlike machines, humans can immediately understand when another person is talking about a different thing and can quickly adapt to each other’s concepts. But this is a big problem that machines cannot easily understand each other’s concepts. If machines could be uplifted to a state where they can easily understand each other, many integration problems we experience in the technology stack will go away.
How to Promote Data Accessibility and Integration?
First, agreeing on standardized formats and interfaces will boost the process of discovering and using data from various providers. A more standardized process will make it easier for consumers to integrate new data into their application systems and generate new knowledge.
Another important aspect is to allow variety in how people approach the same problem – but have shared conceptualizations. Sharing how you define a certain concept, how you sampled the data you used, and the activities you did to solve a problem is key in promoting collaboration and easier integration.
“If everyone defines their worldview and makes it available to others, even machines can understand if they are talking about the same thing or not – which will be a great step in the technology we use to model the natural phenomena.”
Previous Episodes featuring the Open Geospatial Consortium
Open Geospatial standards – shared standards to solve shared problems