© fotolia.com/chombosan

Translating Data into Knowledge

Lateral thinking on an interdisciplinary level can be helpful in managing floods of data

Terms such as Industry 4.0 (smart production) and big data are bandied around everywhere these days. Businesses are connecting up their production facilities, capturing data at every stage of the value chain, and even gathering increasing volumes of data on the purchasing behavior of their clients. But now that they have these huge volumes of data and information, what should they do with them? By applying methods and algorithms on an interdisciplinary level, firms can gain new insights into complex relationships and, step by step, improve their processes. How this can actually work in practice is being demonstrated by Dr. Philipp Liedl, Managing Director STASA Steinbeis Angewandte Systemanalyse GmbH.

These days, practically every business operates in what looks like a vast ocean of data. But which data is really useful? Which factors are really relevant and can be used to make reliable predictions about things to come? What is the best way to capture and analyze complex interrelationships in visual terms? What methods can be used to make predictions about future developments? And how can analyzing data help companies make better decisions regarding this strategy?

Many companies, particularly SMEs, feel uncertain about their chances of success with data analysis projects (“data science” initiatives), and this stops them even getting out of the starting blocks. The actual aims of such projects are often captured in nebulous descriptions or people simply do not even understand them. One problem is that managers are sometimes inexperienced, but they also lack the right vision regarding the doors data science projects could open to them, or how to make effective use of business data. Another issue with data is that it is often unstructured or has to be pulled together from different sources, so a first step would have to be to painstakingly prepare data. Also, these days analyzing data or implementing such projects is not part of everyday business, so resources have to be freed up and new priorities need to be set. STASA Steinbeis Angewandte Systemanalyse GmbH has been tackling such issues for over 20 years. One of its strengths is its ability to think laterally, taking statistical methods originally developed and used in other fields and transferring them to new areas of application in order to come up with new ways to solve data science projects.

Often, using data analysis techniques or self-learning algorithms can result in even more questions than answers. For projects in this area to succeed, it makes sense to involve experts with an understanding of each field of application. This ensures there is sufficient transparency regarding the interplay and relationships between different factors.

Over the last couple of years, STASA has been developing examples of best practice and showing its clients in trade and industry, as well as public administration, how to implement projects, successfully analyze data, use modeling, or create forecasts. The aim of the Steinbeis Enterprise is always to work in close collaboration with its clients in order to take the specific idiosyncrasies of each area of application into account, as well as related solutions. This allows the Steinbeis experts to work on the data science projects for and with its clients in ways that ensure it is a win-win situation for everybody. Using different data analysis and modeling instruments properly, including across different specialist areas, is a good way to unveil new opportunities to apply methods at an interdisciplinary level. One example of this is STASA QC software, which helps optimize machine phasing and quality forecasts in manufacturing processes. The software is based on self-learning and self-structuring models, and it works by drawing on the specialist knowledge of the user regarding a specific production process. It teaches itself how to work out different relationships by looking at correlations – either between the machine settings and sensor data (or process readings), or between the machine settings and different quality assurance measurements, machine cycle times, or energy requirements. The software is marketed worldwide by Kistler Instrumente AG from Winterthur in Switzerland. Kistler is a leading manufacturer of piezoelectric pressure sensors used to monitor processes found in plastic injection molding. Close collaboration with the company allows the know-how of STASA to flow successfully into new analytical functions offered by Kistler’s hardware systems, as well as its software assistants.

Another example of successful application is a joint project with the Reutlingen-based institute Dr. Foerster GmbH & Co. KG. The project involved merging physical models and pattern recognition processes to develop an algorithm for automatically detecting unexploded bombs. The system examines data gathered from magnetic readings of large areas of land and determines the exact location, depth, and diameter of bombs. This was the sort of task previously carried out manually by experts involving huge time investments. By combining the team’s algorithm with a filtering technique used in image processing, the area of technology application could be extended to archaeology. Now, even before investing major resources in digging up artifacts, objects can be quickly identified, despite the fact that – compared to unexploded bombs – archaeological artifacts tend to be bigger and thus more difficult to identify using magnetic fields. The project was awarded the 2011 Steinbeis Foundation Transfer Award – the Lohn Award.

STASA has also developed a tool called the Regional Change Monitor (RCM), which has succeeded in using population movements in local areas (people moving house) to understand the attractiveness of different regions and certain interrelationships, simply by processing data and making it more understandable. In a joint project with IW-Consult GmbH (a wholly owned subsidiary of the Cologne Institute for Economic Research), the Steinbeis Enterprise offers methods for analyzing the strengths and weaknesses of rural districts, cities, and communities, using this to derive recommended courses of action. The RCM is based on an established migration model developed by Weidlich and Haag. The model draws on a master equation that has been used for years in physics to understand statistical changes of state (Haag, G: Modelling with the Master Equation. Solution Methods and Applications in Social and Natural Sciences. Springer Publishing, 2017).

STASA has conducted a number of other successful interdisciplinary projects recognizing anomalies in time series, such as an evaluation of test bed data (condition monitoring), local population forecasting, assessment of regional indicators, analysis of key indicators used in the healthcare industry, and evaluations of traffic and locations.

With many projects, success is not just about gathering the large volumes of data. The ideal approach is often to take things slowly in small steps in order to work out the most effective way to spot interdependencies. This fuels new ideas regarding different options for using business data. Newly acquired knowledge can then be expanded step by step, adding more data sources to improve solutions. STASA uses its successful approach to generate new knowledge and thus add value for its customers. Maybe it’s time for you or your business to do the same. Transform your data into knowledge. The STASA experts are lateral thinkers and happy to help.

Contact

Dr. Philipp Liedl is managing director of STASA Steinbeis Angewandte Systemanalyse GmbH. The aim of his enterprise is to introduce methods into business for systematically analyzing data. The focus of his work lies in data analysis itself, quality assurance, the optimization and control of production processes, and various projects related to urban and regional development, location analysis, and transportation development.

Dr. Philipp Liedl
STASA Steinbeis Angewandte Systemanalyse GmbH (Stuttgart)