Eight expert speakers from different industries made Data Insights Day in our Dornbirn offices an event that was insightful and educational. After a short introduction by Eva Schönleitner, CEO of Crate.io, and Christian Lutz, Founder and President of Crate.io, the event kicked-off and was moderated by Guntram Bechtold from StarsMedia.
Key take-aways from the eight speakers:
- Predictive models are getting better and better but it is also important not only to focus on that but to deliver the right data to the right person in the process to make their work easier
- Tampered data puts businesses at risk, so trustworthy data is key and their are ways to make sure that data is not tampered
- Benchmarks are still the best way to find the right database for each use case. Since the database market is big and diverse benchmarks as a service can be a good option
- The health impact of individual shopping baskets can encourage people to have a healthier lifestyle
- Bringing machines like robots fully to the web makes it easy for people to make use of them. But it also helps machines making use of other machines autonomously which will be a radical change
- It often depends on a single person to make radical innovation possible
- Also small machine learning applications can make the work of people much easier and reduce the amout of repetitive work for them significantly so that they can focus on other tasks
- Collecting data from various sources, transforming it in a generic format, moving to a central data storage and enrich it with topology information to provide it securley are the key components to make intralogistic data projects a success
- Real-time data insights become more and more important in a high-speed production environment with lines producing 120.000 pieces per hour where 1 line has more than 1 million sensors
At this invite-only event, the audience not only heard about many different aspects of data, but could also ask questions and discuss with the speakers about the current state and the future of data insights.
Below you can find some pictures from the event and the abstracts of the talks. Click on the button below to watch the recordings of the talks.
Upcoming Data Insights Day
If you want to be part of the next Data Insights Day, please fill out the form and we will inform you about our next event.
What to expect?
Identify leading trends and business opportunities to scale your data strategy during Crate.io's Data Insights Day.
Crate.io brings together regional thought leaders, customers, and partners to equip you with the technical know-how, practical advice, and expert insights on connecting data insights to business outcomes.
Request Data Insights Day
CIO Rauch Fruchtsäfte
In a system-integrated "production world", available data is very often hoarded but not actively used. One of the main problems in integrating this data is the hybrid system world that needs to be connected, as well as the technical challenges with regard to the quantity, structure, and temporal processing of the data that accumulates. In the topic "DDQA - Data-Driven Quality Assurance" we deal with the use of existing data regarding "real-time" monitoring of production, & quality feedback with special consideration of the necessary IT architecture.
The construction industry accounts for 50% of material resources taken from nature, 40% of energy consumption, and 50% of total waste generated. To minimize these numbers, there is an ongoing shift in the industry towards sustainability, industrialization, and automatization. One approach to tackle these challenges is through the lens of data. Recent studies indicate that 95% of data generated in the construction industry are not used. This talk shows the data utilization efforts in the Rhomberg group: how we try to make our company more data-driven and build intelligent systems to support the core business.
Today's warehouses are complex systems with a very high automation degree. Interconnected conveyor systems on various levels, racks with automated storage and retrieval technologies like shuttle systems, lifts and mini-load cranes, highly optimized manual workstations and fully automated loading, un-loading, packing, picking workstations with stationary robots and mobile robots roaming on the shop floor. These mechatronic devices are controlled by distributed software components for controls, material flow and warehouse management.
The key to successful operation of these warehouses lies in having a holistic view on the entire system based on data from various components like sensors, PLCs, embedded controllers and the involved software systems. All these components can be seen as “data silos” distributed across the entire site - each of them storing just some pieces of information in various data structures and different ways to access it.
This session is about how this data can be retrieved, transformed, stored in a central place and how to link it with the static site topology so that useful information can be provided. The session will cover how to make this information understandable and accessible for inhouse experts and business customers. Some applications will be explained that provide insights on the availability and conditions of the mechatronic equipment as well as the performance of logistic processes.
In this talk, Simon will give an overview of two current research projects that are being conducted at the Institute of Computer Science of the University of St.Gallen. Both projects target how we might increase the accessibility of data, but target almost completely disconnected spheres: Human Nutrition and Industrial Automation. Furthermore, in the course of the talk, Simon will introduce the new School of Computer Science at the University of St.Gallen and its research programs.
Why are so many companies great at incremental innovation, but at the same time fail badly with radical innovation – and how can this be changed? In over 60 personal interviews covering more than 20 large industrial companies in Switzerland, Austria and Germany, we took a deep dive into these questions and found a surprisingly consistent picture of blockers and success factors. In my presentation, I'll give a first peek into the findings of our interviews, the pattern that led to success and the underlying dynamics.
Designing efficient IT systems for data-intensive and highly scalable application domains is complex and risky. Why? Because information and data points are often missing for an informed, holistic decision. For data-intensive applications, the database and cloud (or on-prem) resources are the critical components - and performance and cost are the key SLAs. In this talk, we share our experience and show how this process can be simplified and made more efficient through cloud database benchmarking. Using a real world IoT project from the automotive industry, we demonstrate how to find the optimal cost/performance trade-off for operating a database in the cloud and how to assess performance-related SLAs on multiple levels.
Demand forecasting and planning is one of the most important business use cases of applied AI and predictive analytics. At the same time, many companies use a spreadsheet-based approach in their planning process. In this talk, Franziskos will show various examples of where demand forecasting and planning are important in industries such as retail and supply chain, how to do it properly and how the value of machine learning shows when it comes to predictive accuracy.
Imagine the impact on your business when you make decisions based on tampered data. What happens if attackers hijack smart sensors or other critical devices and send fake data to manipulate decisions made by operators, data models or applications? Across almost all industries, the growing number of IoT devices, increasing interoperability between systems and data transfer across companies continuously attracts more and more attackers to not only steal or encrypt data but also actively tamper it. This “industrial fake news” has the potential to cause even more damage than the fake news on social media. Data notarization provides a new level of data security from inception to reduce blind faith in data and establishes more credible trust from sensor to consumer - even when shared across companies.