On this episode, hosts Phil Seboa and Ed Fuentes explore the world of industrial data engineering with Keiran Stokes, Director and Head of Technology at Thred. They discussion overcoming data sharing challenges in industrial systems, the critical importance of context in data analytics, and the innovative potential of open-source software.
Overcoming Data Sharing Challenges in Industrial Systems
One of the pressing issues Keiran Stokes addresses is the complexity of data sharing in industrial systems. Unlike IT systems designed for data movement, industrial systems focus primarily on command and control, leading to significant integration challenges. “Industrial systems are not designed to easily share data like IT systems. This is why we see a lot more challenges when we try to integrate them together,” notes Keiran.
This discrepancy often results in siloed operations, where valuable data is trapped within individual systems, hindering overall operational efficiency. To tackle this, Keiran emphasizes the development of data engineering solutions tailored to bridge the gap between industrial and IT systems. These solutions are designed to promote seamless data exchange, ultimately enhancing operational efficiency and driving insightful decision-making.
For industrial businesses struggling with these issues, Keiran advises starting with a clear strategy. Begin by identifying specific integration points and utilizing specialized tools to facilitate data sharing. By addressing this challenge head-on, companies can unlock the true potential of their industrial data.
The Power of Context in Data
Understanding and leveraging context is essential in transforming raw data into actionable insights. Phil points out, "Contextualizing raw data is challenging because operations are often siloed. We end up with massive datasets that lack proper context, making it hard to derive meaningful insights."
Keiran elaborates on the concept of knowledge graphs—a flexible data structure consisting of nodes and relationships that better represent the interconnected nature of industrial data points. “Knowledge graphs are used to store and codify context within Thred's systems,” explains Keiran. This approach allows businesses to enrich their data analytics by preserving the relationships and processes that add meaning to raw data.
Moreover, the integration of Large Language Models (LLMs) plays a crucial role in this context. Keiran describes how LLMs help understand and structure complex datasets, making it easier for organizations to analyze and derive meaningful insights. By combining knowledge graphs and LLMs, businesses can more effectively utilize their data, turning it into a powerful tool for strategic decision-making.
The practical application of this approach is evident in Thred's new data contextualization tool, "3 Cloud." While still in testing with a few select customers, this tool aims to model factory data, integrate it into a company's IT data stack, and handle all transformations on the IT side. This integration simplifies data management and enhances the overall quality of insights derived from industrial data.
Embracing the Potential of Open-Source Software
Open-source software presents a significant opportunity for innovation in the industrial sector. Keiran emphasizes, "Open-source software allows commercial companies to lower costs and add value through additional services like infrastructure, hosting, and support."
Adopting open-source solutions can offer several advantages, including cost savings, increased flexibility, and accelerated innovation. These solutions also foster a collaborative environment where continuous improvements and advancements are encouraged. Companies like Timescale, Portainer, and DBT are prime examples of open-source technologies that have been successfully commercialized, demonstrating their viability and benefits.
Phil adds, "The collaboration between open-source and commercial products enhances innovation and scalability."
For industrial businesses, this means that embracing open-source technologies can provide a competitive edge. By opting for open-source solutions, companies can avoid the limitations and costs associated with proprietary systems, thereby fostering a more innovative and adaptable technological environment. This approach not only drives down expenses but also increases the capacity for scalability, making it a win-win for industrial enterprises.
Key Quotes From The Episode
Keiran Stokes emphasizes, "Industrial systems are not designed to easily share data like IT systems. This is why we see a lot more challenges when we try to integrate them together."
Phil Seboa adds, "Contextualizing raw data is challenging because operations are often siloed. We end up with massive datasets that lack proper context, making it hard to derive meaningful insights."
Key Takeaways
- Data Sharing Challenges: Industrial systems struggle with data integration due to their design focus on command and control rather than data exchange.
- Importance of Context: Contextualizing raw data through knowledge graphs and leveraging LLMs is crucial for extracting meaningful insights from industrial data.
- Open-Source Potential: Embracing open-source software can drive innovation, reduce costs, and enhance scalability for industrial businesses.
Wrap Up
In this episode, Keiran Stokes discussed the complexities of data sharing within industrial systems, stressing the need for solutions that facilitate seamless data exchange. He underscored the power of context in data analytics, explaining how knowledge graphs and LLMs can transform raw data into valuable insights. Lastly, the conversation highlighted the potential of open-source software in driving innovation and reducing costs for industrial businesses.
These insights are crucial for companies looking to elevate their data engineering practices. By focusing on integrating and contextualizing data, and exploring open-source solutions, industrial businesses can unlock the full potential of their data, driving efficiency and fostering innovation.
To address these challenges, consider forming a dedicated team to assess and optimize your data sharing processes. Leverage specialized tools to integrate and contextualize your data, ensuring it provides actionable insights. Embrace open-source solutions to stay cost-effective and innovative, allowing your organization to adapt and scale in a rapidly changing industrial landscape.
About the Guest
Keiran Stokes is a leader in IT/OT integration and data strategy, renowned for his expertise in converging operational and information technologies to drive digital transformation. With a strong focus on supporting New Zealand’s industrial sector, Keiran specialises in enterprise and data architecture, helping businesses unlock productivity through digital solutions. His passion for bridging the gap between the factory floor and the cloud has positioned him as a thought leader in operational data, as well as in the engineering and operations that transform that data into insights and actionable improvements.
Connect with Keiran on LinkedIn:
https://www.linkedin.com/in/keiran-stokes/
Stay tuned for more expert insights on Unplugged: An IIoT Podcast.
