Katana –
Big Data Analytics


What needs to be optimized?

Working interactively with you and your experts, we come up with a use case for individual improvements. During an initial discussion, we determine which errors or faults need to be identified from a data perspective and where there is room for improvement. Our data scientists and your experts then hold further discussions during which they develop and enhance their examination and assessment of the data. The result of this important first step is the business value definition of the use case.


  • Exploration
  • Business value definition
  • Data identification


Does it work?

Our data scientists prepare your data specially for your use case. Our key area of expertise lies in developing an individual algorithm, e.g. from machine learning or cognitive intelligence. Patterns, irregularities and outliers can therefore be identified and previously unknown causal links are revealed. The basis for optimization is now in place.


  • Visualization
  • Algorithm development
  • Validation

Your raw data ...

Structured sensor data from an industrial machine

Data from sensor measurements are usually collected in the form of time series. Because sensor data are often subject to interference and incorrect, they need to be preprocessed before they can be analyzed further. This includes for example finding and eliminating measurement and transmission errors. In order to pre-process your data optimally, we develop tailor-made solutions and use standardized approaches.

Semi-structured log data from a web server

Both machines and IT systems describe status changes by outputting discrete events in the form of log data. These are generally semi-structured and are difficult to analyze owing to the huge quantity of information and the frequency at which they are output.

Entry ticket to the USU ITSM Valuemation suite

80% of company-related data are unstructured. These include texts containing valuable hidden knowledge which is overlooked without intelligent text-analysis procedures. With the help of statistical and linguistic measures, you can obtain structured information from these data, which is useful during further analyses.

...become our analyses

Automatic measurement correction in the event of sensor faults

Automatic measurement correction in the event of sensor faults
Motor torque over time recorded during test runs with outliers and drift (blue), with outliers removed (green) and with outliers and drift removed (red).

If sensors suffer temporary faults when taking a measurement (see blue curve in the defective section), this can often render the entire measurement useless. Because taking measurements can be a very time-consuming and costly process, these series of measurements should not be discarded if possible.

With the help of solutions developed by us, defective measurement intervals can be corrected (see blue and green curve in the defective section) and measurements can be adjusted for drift (see red curve).

Real-time identification of outliers in sensor data

Snapshot of outlier identification in real-time sensor data. Measurements outside the tolerance range (light blue) are identified as outliers (red) and removed. The moving median is inserted in place of the outlier.

Identifying and eliminating outliers during operation is extremely important, especially for real-time analysis systems. This helps to prevent errors caused by outliers being reproduced in subsequent processing steps and thus falsifying analysis results.

Unlike the classic cluster analysis method, our solution allows outliers in sensor data to be both identified and eliminated during operation.

Adaptive anomaly recognition in real time

Adaptive anomaly recognition in real time
Recognizing anomalies in machine data with low (green) and high (yellow) tolerances. If a value lies outside the green area, this could constitute a warning. If a value lies outside the yellow area, an alarm could be triggered.

Anomalies are irregularities in the measurement data from machines and therefore play a key role in identifying faults.

With the help of the anomaly recognition system developed by us, historical data are used to automatically identify various machine operating phases and to calculate phase-specific tolerance thresholds. The sensitivity of the anomaly recognition system can be set by the user on an individual basis. As a result, even slight deviations from normal machine behavior can be identified in real time while the machine is operating.

Predictive Maintenance

Predictive Maintenance
Comparison of a standard prediction (green) with the prediction developed by us (red). The red area represents the confidence of our prediction, while the blue curve shows the actual course.

In addition to real-time anomaly recognition in sensor data, a long-term prediction of the development of relevant key variables over time plays a central role as well. Key variables could be, for example, a machine’s performance or oil quality. If the development of these key variables over time can be predicted with sufficient accuracy, spare parts can be ordered on time and service technicians can be scheduled for maintenance. This helps to prevent costly downtimes.

Prediction approaches based on a linear extrapolation of measured variables can lead to highly inaccurate predictions. The approaches developed by us are based on machine learning procedures. They offer improved predictions with additional information regarding the uncertainty of the prediction. See illustration.

Fault/cause analysis

The event tree shows the links between the fault events of a machine. Read from left to right, selected events are matched to potential trigger events. The colors indicate the likelihood of a link between the relevant events. Filled in nodes are not matched to any trigger events.

Both complex machines and IT systems describe their status by outputting discrete events in the form of log data. These generally contain important information regarding the status of the systems and developments over time. Thanks to self-learning procedures, our solutions are able to structure log data and identify the most likely causes of the occurrences of sudden events. As a result, we reveal any relevant links.


Now the practical part!

Transformation to the Katana platform now takes place. In order to do this, our development experts migrate the prototype algorithm so that it can be used effectively and reliably with large quantities of data. The subsequent integration into your target environment allows permanent further development and optimization.


  • Migration
  • Integration
  • Maintenance


Hit the market with a new portfolio!

With this approach, your data-driven portfolio benefits from two smart services. “Predictive maintenance” is used to increase productivity through proactive recommendations, while peer group comparisons result in a performance-increasing consultancy approach that optimizes overall system effectiveness. Smart services ensure that you stay ahead of the competition.


  • Smart services
  • Business value
  • Market position