Data analysis - a prerequisite for the smart industry
The real benefit behind concepts such as AI and IoT is realized by transforming data into insight. Successful data analysis is based on structured collection, storage and use of information to generate business benefits, innovative power and business development.
Automation Region has spoken to three experts - how do they view developments in the field and what advice do they want to give organizations that want to get started creating value with the help of data analysis?
Great potential values
Kristian Sandström is a senior researcher at Rise in Västerås and he runs several applied industrial research projects where data analysis is an important component.
– Many of our projects use data analysis with the aim of developing solutions in AI or machine learning, but already on the way there, basic processing, analysis, and visualization of data usually generate new insights, questions and sometimes unexpected answers, says Kristian Sandström.
He also emphasizes that data analysis can create value in different ways - for example by describing a current situation, optimizing businesses and processes, or to create innovation. The analysis provides a common and objective picture of the business around which all functions within the organization can be gathered. At the same time, good domain knowledge and a certain caution in handling are required.
– Use the right processes and methodology for the purpose, says Kristian Sandström. What is the goal, which case should be addressed with data analysis and how is success measured? Who are the users and how should they absorb the results?
Predictive process control and fact-based decisions
Micael Baudin, Senior Development Engineer at Seco Tools, and Fredrik Jakobsson, responsible for AR & 3D at Softgear, have collaborated in the development of a visualization solution for the industry. The purpose of the project is to make complex data available to operators and strengthen their ability to assimilate information to make diagnoses and troubleshoot predictively.
– We need to move from reactive process control to a more predictive approach throughout the product's life cycle, says Micael Baudin. Then we can achieve more precise tolerances, reduced lead times, higher machine utilization, lower energy consumption and lower maintenance costs. Which in turn gives our customers added value on several levels.
– Basically, data analysis is about compiling relevant information and ensuring data quality, says Fredrik Jakobsson. We need to make the information more accessible and visual to facilitate fact-based decisions.
Micael Baudin highlights two aspects that he believes are very important when it comes to data analysis - ensure a constant data flow in several stages of buffering, and ensure time synchronization for all data streams.
– The demands on the data sampling frequency are increasing and many connected devices lack functionality for time concepts, says Micael Baudin. A data analysis will never be better than the data on which the analysis is based.
Fredrik Jakobsson emphasizes the importance of knowing where the data comes from and that there are routines to ensure that the information is correct. He also underlines that it is important to present the analysis in a good way.
– The results must be made available in an equal and democratic way through pedagogical and well-thought-out interfaces that provide increased benefit from the analyzes, says Fredrik Jakobsson.
How to create value with data analysis
- Adapt processes and methodology for the purpose - value-creating data analysis requires domain knowledge and accuracy.
- Strive for predictive process control throughout the product's life cycle.
- Make sure you know where your data comes from and that the information is correct.
- Provide information and results with well-thought-out visualizations.
Automation Region continues to work with data analysis
Industrially applied data analysis is fundamental for two of Automation Region's thematic areas - connected industry and flexible automation. In digital projects, the challenge lies in extracting data from organizational downpipes and collecting them in a standardized and accessible form.
Technically, it is about collecting, processing and visualizing data in a way that creates value for the organization. But with increased use of digital tools, the organization needs to adapt to the new conditions. Surveys show that more than half of all digitization projects fail and a common reason is that the organization tries to apply new digital tools to old structures.
In 2020, Automation Region's research and innovation group will focus in particular on data analysis as an enabler for smart industry and increased innovation. If you want to be part of that work or participate in research projects in these areas, you are welcome to contact the group's chairman Anders Aabakken or Daniel Boqvist, program manager for research and innovation at Automation Region.