Big Data & Data Science

We know how.

Boundless possibilities.

To use big data effectively requires new and innovative forms of information processing, including unconventional data management technologies, architectures and analytical functions. The potential application scenarios for big data are meanwhile both extremely comprehensive and very individual. But regardless of the particular use case, the same implementation challenges arise time and time again. It always comes down to finding a way to efficiently handle large data sets, quickly process the data streams, and master complex analyses. That is why we have developed implementation solutions for big data projects which address these three challenges.


  • Store & process very large data sets
  • Capture & process fast data streams
  • Advanced Analytics & Data Lab

“Trivadis masters in equal measure both traditional and new big data technologies, as well as complex analytics. That is the telling advantage.”

Big Data Canvas

We take the Big Data Canvas architecture approach to create the full scope of a big data solution using seven coarse-grained architecture modules. Each of these building blocks forms a functional area for processing analytical information, which is initially broken down by fine-grained architecture building blocks, before being finally stored by the corresponding solution components, so-called solution building blocks. This approach allows the realization of highly standardized and integrated big data architectures to accommodate every individual business case.

  • Standardized architecture
  • Proven solution components
  • Investment security

trivadis big data science

Store & Process Very Large Datasets – Enterprise Data Hub

The enterprise data hub is the central point for storing all business-relevant data in a comprehensive operational data store (ODS). All the data are available in their original state, and are continuously collated and saved in the greatest detail. This all takes place in an extremely cost-effective manner coupled with adequate computing capacity either for

running analyses directly on the data hub platform, or as a central repository for all the other IT systems. A consistent backup strategy for all the historical data allows detailed data to be historically accessed in full even years later, or error analyses to be run with unprecedented reliability.

Full data historization!

The main problem with historical data is not their volume, but rather their heterogeneous nature. Structures and significances are constantly changing. The big data lambda architecture allows for a uniform view of historical data – regardless of their structure. And brand-new structural changes will never prevent data from being recorded either. Nothing is lost.

  • Standardized architecture
  • Historical data is always accessible
  • Equipped for future analysis methods

trivadis store process

Speedy data streams - real-time analytics

It is becomingly increasingly crucial to be able to analyze data streams in real-time, not only from social media sources, for instance, but above all from sensor and telematics systems. However, the parallelism of multiple simultaneous data streams and the frequently changing and uncontrollable data rates represent major challenges in processing such data.

Our approach is based on a flexible lambda architecture, which we configure to accommodate the customer's requirements and implement using proven technology components from the Hadoop ecosystem.

The architecture is crucial!

The architecture as success factor. The lambda architecture we propagate masters the major challenges associated with processing real-time data streams. This architecture can be scaled to cope with any number of data streams as well as the transmitted data sets. Components can be replicated to boost the system's performance accordingly. Processing components for real-time analyses can be incorporated in the data stream.

  • Standardized architecture
  • Scalability of the overall solution
  • Ready for real-time analytics

trivadis speedy data streams

Data Lab – advanced analytics in a sandbox

Power users, BI analysts or data scientists need more than just access to the data warehouse and BI tool. What are also required are powerful analysis platforms which allow heterogeneous data from different sources to be analyzed in self-service mode.

We can design and build you the perfect workplace for exploratory analyses. This means that you gain new knowledge without week-long lead times.

Everything under your control!

Depending on your requirements, we link existing DWH data with other information sources such as social media, log files, sensor data or documents. The right technological approach means that the system as a whole delivers the best possible performance. Whether different systems are connected virtually, or data are physically integrated at a suitable location, is dependent upon the required performance and data quality. We will apply our technical and professional expertise to develop a suitable solution that perfectly meets your cost-effectiveness expectations.

  • Unified Analysis
  • Advanced Analytics
  • Semantic Web

trivadis data lab

Trivadis, Peter Welker, Senior Principal Consultant

Peter Welker

Trivadis, Senior Principal Consultant