Empowering domain experts to work with data

The Swedish startup Viking Analytics has developed an industrial analytics solution specifically targeted to asset-heavy companies in the chemical industry, enabling them to analyze, monitor and predict the performance of their assets and processes. In this interview, Dr. Rajet Krishnan, co-founder and CTO of the startup, which is part of 5-HT’s digital ecosystem, explains how customers from chemical industry can benefit from using their software. Dr. Krishnan highlights, “What makes us unique is our ability to capture expert knowledge and to build smart applications on top of that.”

What is the story behind the foundation of Viking Analytics?

My co-founders and I saw that there was a huge need for advanced data analytics in heavy industry, especially in chemical industry. Viking Analytics was founded in 2017 by Stefan Lagerkvist, Arash Toyser, Johan Isaksson and me. At that time, we were all running our own consulting companies. While on a consulting assignment for the Dutch multinational company AkzoNobel, we started realizing that one of their biggest challenges was dark data: they were collecting a lot of data without doing anything about it. In chemical companies, there are often not enough domain experts who are able to analyze and make sense of the collected data, and there are also not enough data scientists or programmers inhouse to develop machine learning solutions themselves. These findings formed the foundation of Viking Analytics: we wanted to empower domain experts, who are not necessarily data scientists, to work with data in an efficient way.

How do your enable domain experts without advanced data knowledge to make use of their data?

With our software solution, domain experts don’t need to have any programming skills or advanced data knowledge. Our AI tools are specifically designed to empower domain experts to gain insight from their data without the support of a data analytics team. Especially the user interface makes it very convenient for our clients to work with our software.

Who are your customers?

We have a huge spectrum of customer segments within chemical industry. One focus are asset-heavy companies, as our AI solution allows to analyze, monitor and predict the performance of various assets. Furthermore, we focus on companies that are trying to automate their processes. Based on data collected by sensors, we provide automation solutions. We have a strong, growing customer base distributed in different European countries like Sweden, Austria, the Netherlands and Germany. For example, we are currently supporting AkzoNobel in an energy optimization project, and we are cooperating with Borealis AG to improve their root cause analysis. We are also working with one of the largest machine manufacturing companies in Germany in the area of predictive maintenance.

What applications are there for your software solutions, and how do your customers benefit from using it?

Usually, subject matter experts start with raw data collected from assets and processes, which is cleaned, prepared and processed to get useful information. Information is then analyzed by the experts in order to reach actionable insights. Finally, insights are put to action by means of (expert-annotated) data-driven applications deployed in real-time. The level of involvement of the subject matter experts in the above value chain dictates the success of the deployed application. Deep and efficient involvement in this value chain can be realized with the broad spectrum of features in our tool MultiViz.

Our solution allows to quickly connect, clean and prepare the data so that experts can use their time to do what really matters: understand what events and abnormalities happened and why they happened, identify operational modes, failure modes and other abnormal behavior. If someone needs to spend hours just looking through spreadsheets or manually preparing data, then it is natural that there will not be much time available to capture the knowledge in a way that it becomes available for collaboration. With our tools, it shouldn’t take more than a couple of clicks to be able to have a holistic context around the different anomalies discovered. Combining the anomaly summaries, data from multiple databases and expert annotations in a contextualized way translates into valuable knowledge that can be either shared via reports or centrally saved for easy and quick consultation. These insights also form the basis of real-time smart manufacturing applications like process optimization, anomaly detection and predictive maintenance; optimization that can be deployed at ease with MultiViz.

What sets you apart from existing industrial analytics solutions?

There are many powerful industrial analytics solutions on the market. However, we stand out in terms of how capable we are to capture experts’ knowledge and to build smart applications on top of that. When we started with Viking Analytics, we soon realized that domain experts had a lot of valuable knowledge about their data, and we wondered how we could enable them to transfer this knowledge into the software and to operationalize their insights. Therefore, combining algorithms with expert knowledge is the core of our software. We also have a university collaboration on researching algorithms for optimizing and automating capturing expert feedback. Our software solution enables experts to discover and understand anomalies and events and provides recommendations on how to react. Based on the experts’ feedback, we improve our recommendations. MultiViz is a highly intelligent software that really learns from the experts. Furthermore, with our solution, even large companies can go from raw data to a complete industrial AI solution in a matter of months. Not many competitors can offer this kind of full solution that can be scaled up in such a short time.

What are the next goals for Viking Analytics?

In times of Corona crisis, we are currently trying to understand how we can help companies to make the best out of the situation. Now is the time for investing into digitalization and software solutions, as topics like remote monitoring or remote maintenance are getting more and more important. This year, we are also planning to set up an office in Germany, for which we are currently hiring sales and technical employees.

What do you hope to achieve by being part of our Digital Hub?

The German market is very important to us for a multitude of reasons: the large heavy industrial base including chemical companies and the many hidden champions in Germany. These champions are highly specialized companies that are market leaders in their specific field. If they are looking for a digital solution, they can either build their own team or work with a strong partner like us. Being part of 5-HT is a great opportunity for startups like us who are interested in collaborating with German companies. We are looking forward to building fruitful cooperation and enabling German companies to come up with digital solutions within just a few months. For us, being successful in Germany is an important starting point for expanding our business globally.