With data science, you can identify undiscovered potential, recognise patterns and make predictions. After an explorative analysis, you make decisions based on data and can optimise and automate your processes.
Our performance in the area of data science
Our data science experts analyse and explain your data and develop modern solutions together with our experienced software developers. In doing so, we pay attention to data protection compliance. So that you can benefit sustainably from the possibilities of artificial intelligence.
- Object recognition
With the help of computer vision, you can recognise and localise objects and people in images. This can automate processes and simplify workflows.
- Data integration & preparation
- Visual and statistical analysis
We merge and cleanse disparate data sources and help you gain new insights into your business.
- Time series analysis
- Anomaly detection
- Predictive maintenance
- Customer segmentation
We look for groupings and anomalies in your data and use insights gained to make predictions.
Three steps to success
Our approach to data science projects:
During a personal meeting, we identify data potentials in your business model.
With an explorative analysis and a PoC, you get an insight into the quality and possibilities of your data in the shortest possible time. Depending on your needs, you will receive a written report or a visualisation.
We make your software intelligent: depending on your requirements, we integrate an intelligent algorithm into an existing application or develop a completely new software. If required, we can bring your application into the cloud.
We take data protection seriously. However, to give you an insight into our expertise, we use free data sets to show you our procedure and what conclusions can be drawn from the respective data.
Explainability of machine learning algorithms
Many machine learning methods are so-called black-box algorithms that do not offer direct access to what is learned. This means that it is initially unclear whether a calculated prediction can be trusted. This is where tools for the explainability of machine learning algorithms such as LIME or SHAP can help. In our showcase, we show how you can use them to debug and improve the process from data collection to prediction. Learn more
Clustering & anomaly detection
On the basis of our Open Food Facts Showcase, we show you how you can use clustering algorithms to detect naturally occurring groupings as well as rare observations - so-called anomalies. The latter can be faulty user entries of an app or rare products, but in the worse case also fraud cases or server attacks. This gives you deep insights into the structure of your data and allows you to use this knowledge to make better decisions. Learn more
Using neural networks to combat microplastics in water
Most sewage treatment plants are not equipped with a so-called fourth purification stage. This can help to remove pharmaceutical residues and tiny particles such as microplastics from the treated water before it is returned to the water bodies. Since they are expensive to purchase and operate, it is advantageous to know beforehand how high the load actually is. In the interdisciplinary research project VAMINAP, artificial neural networks help to detect pollutants. Learn more
Focus on the customer with data science analyses
Using historical data on sales figures and customer activity, you can gain deep insight into your business processes through data science analyses. Being informed about customer preferences and trends also means saving expensive storage space while guaranteeing on-time delivery. We show how this works with our e-commerce showcases.