Data modeling is a crucial process for the modern production company as it transforms raw production data into standardized, coherent, and usable information.
Therefore, data modeling is the key to advanced analysis and optimizations as it consolidates the many data sources of production with different measurement systems and sampling frequencies into one data model where you can compare and utilize your data.
It provides a deeper understanding of your production, identifying optimization opportunities, and improving your decision-making.
What is Data Modeling?
Data modeling involves moving and structuring the vast amount of production data available so it appears standardized and related to each other—typically in one integrated cloud-based data model. This includes data and attributes from the entire production line, whether it is real-time data from your machines, recipes, ERP system, or OEE data.
A central part of data modeling is establishing traceability. Many associate traceability with the ability to, for example, recall products if a problem arises. However, traceability is also key to linking data across production, preferably automatically and in real time.
Data Classification
With data modeling, you can classify data and information in an entirely new way. Therefore, data modeling is a crucial element in a digitized and efficient production as it creates a standardized basis for analysis and valid data processing.
In broad terms, you can view it like this:
With an organized data model, you move from having data to obtaining information. You go from individual data points with different sampling frequencies to holistic views and a comprehensive overview of your entire production.
This is beneficial for traceability, ESG reporting, and overall production efficiency.
Data Model vs. Data Modeling
Data modeling is the path you follow to create the structured plan—the data model itself—that describes how your data should be organized and used in your applications and databases. So, a data model is your blueprint for data management and the result of the entire data modeling process.
Why is Data Modeling Important for the Production Company?
Data modeling is a crucial discipline in digitization for production companies as it provides competitive advantages such as advanced optimization opportunities.
Once you have organized production data, you can perform complex data analyzes with many variables and using tools such as Power BI, AI, and Machine Learning.
The output is optimization opportunities across the entire production that you would not have seen in a more fragmented and siloed data structure.
What is the Production Company’s Data Challenge?
At integra2r, we consider data modeling as the next step on the digitization ladder, particularly for production companies that are already good at collecting data, which most are.
This also applies to companies that have optimized production but wish to take their production optimization to the next level.
Many of the production companies we collaborate with already have good and valid OEE data, data from the quality system, maintenance system, and recipe system. They have large amounts of data from the entire supply chain, but few have used data modeling to standardize and relate data to each other and thus gain the holistic production picture.
The challenge lies in different sampling frequencies.
The challenge is that data input comes in many different forms, models, and time intervals from applications, machines, measuring units, systems, and operators. Some systems store data in one way, while others provide data every second. Others provide data every hour.
To convert data into information and usable knowledge, a specialist with in-depth knowledge of each system or site is typically required, which is both resource-intensive and time-consuming.
With a data model, you can see optimization opportunities across your production immediately. And therein lies the great potential and competitive advantage.
Benefits of Data Modeling for the Production Company
The benefits of data modeling are many, but here are three selected benefits for the production company.
Data Accessibility
With data modeling, you make your data accessible in one data model. This is where you transform individual data points into holistic views, turning data into information and usable knowledge that you can implement in your production.
Artificial Intelligence, Machine Learning & Business Intelligence
By properly organizing your data, you can connect advanced AI technologies, Machine Learning, and Business Intelligence to your data model to conduct complex analyses.
You can also link information together using traceability, thus seeing correlations across production in a completely new and far more efficient way than before.
This makes it possible, for example, to extract actionable insights and predictions and transform large datasets into manageable reports and dashboards.
Overall, it can provide a more proactive approach to production optimization, where you can react before problems arise.
Cloud-based Database
The data model can be stored in the cloud, providing a range of benefits for the production company:
1. We can store, extract, and adjust data in large quantities across production sites without affecting production itself.
2. We get IT tools that enable us to store and process virtually all the data we need.
3. It is relatively inexpensive.
Examples of Successful Data Modeling
Below, we have gathered two specific examples of how a successful implementation of data modeling can unfold in improved traceability in terms of resource consumption and real-time dashboards.
Optimized Energy Consumption for ESG Reporting
When you gather data sources from your energy consumption, machines, and production output into one data model, you can track how your energy consumption varies in the production flow of each product.
This provides you with ESG data and a basis for comparison where you can benchmark your energy consumption against your production volume of a single product or an entire product class.
For example, if a type of dairy product requires 10 percent more energy to produce than another, you can choose to optimize that particular production or switch production to a more energy-efficient product.
Real-time Traceability with a Live Data Model
Another approach to data modeling is to look at the Unified Namespace. It is, briefly, about forming and organizing your data in real-time and building a live data model. A central data model that ideally contains all your information in one place, in one structure. This does not necessarily mean that everything has to be redone from the start, as you can start by looking at the most important areas first.
The benefits include, for example, being able to create dashboards that represent the production’s current flow 1:1, thus being able to catch errors before they occur or just as they happen.
In terms of sales and marketing of goods, it also becomes possible for consumers – let’s say in the automotive industry – to follow the production of their new car step-by-step.
You go from being reactive to being able to be proactive and catch errors in the bud, which can save significant costs.
The Connection Between a Manufacturing Execution System (MES) and Data Modeling
Manufacturing Execution Systems (MES) play a crucial role in the digitization of production processes, and data collection is the core of effective MES solutions, where ISA 95 is typically the model standard.
At integra2r, we see the implementation of an MES system as the first – and most crucial – step in working with data modeling. This is because in an MES system, a portion of the data available in production is taken and placed into the system.
This is the first part – organizing data and the foundation for your data modeling.
The challenge is that the MES system does not include all the available data from the start, as an MES system typically only integrates with a portion of the production systems.
At best – and to get the full overview of the production flow – you have an MES system that collects data from the entire production system, but it requires that the system is fully integrated with both the ERP system, the financial system, the quality system, the PLC system, IoT devices, etc. That is, the entire production process.
The implementation of MES and data modeling is a step-by-step process where the MES system is the foundation that you continuously build on – for example, by implementing various cloud solutions or integrations.
Do you have questions about data modeling?
We hope you have gained insight into data modeling and how it can optimize your production.
If you have questions, you are always welcome to reach out to Senior MES Architect Poul Klemmensen at pkl@integra2r.com or (+45) 9154 5445.