Development of a data science infrastructure at AOK Baden-Württemberg
Creation of the optimal framework conditions for the topic of data science using a performant, scalable and at the same time secure IT infrastructure.
Setting up a central infrastructure that meets the heterogeneous requirements of different user types in equal measure.
Develop a hybrid stack of open source and enterprise solutions around existing components such as SAP HANA DB, R and Tableau.
AOK Baden-Württemberg wants to push the topic of data science within the company. The central building block of this initiative is the establishment of an IT infrastructure that is designed for the high speed of development and the agile approach in the field of data science and at the same time meets the high requirements of a productive IT system.
As an AOK subsidiary and specialist for IT services in the healthcare market, ITSCare chose eoda’s data science experts to design a professional IT infrastructure as part of eoda | analytic infrastructure consulting for AOK Baden-Württemberg.
To develop an optimal infrastructure and implementation concept, eoda brought together the data science and IT managers on the customer side as part of an assessment tailored to AOK Baden-Württemberg. The goal of the assessment was to determine the maturity of the existing IT infrastructure and the central data and user stories. The data stories illustrate the path from the data source to the data output to the relevant stakeholder and the associated requirements for the infrastructure.
Due to the large number of different data stories and user types, the central challenge for AOK Baden-Württemberg was to create a central infrastructure solution that optimally takes into account the very heterogeneous requirements.
The assessment identified the following key drivers of potential for the existing IT infrastructure in relation to the requirements in the data science context:
- Expansion of the degree of professionalization
- Improvement of collaborative working and reproducibility
- Increase in scalability
- Modular design with different expansion stages for future challenges (cloud readiness, etc.)
- Integration of core components from the already existing data science landscape
The existing IT infrastructure based on SAP HANA DB, R / RStudio and Tableau already offers Data Scientists numerous tools for implementing even sophisticated models. However, with a centrally anchored concept and the associated professionalization of the IT infrastructure, a technical environment can also be created for the Data Scientists that significantly supports them in their work.
Based on these findings, the experts at eoda have developed an implementation concept for AOK Baden-Württemberg that includes the following areas:
Integration & tools
IDEs for R & Python, use of Docker, etc.
High-performance data access, etc.
Rights & installations
Permissions, package installations, etc.
Reporting & deployment
Reports, dashboards, apps, REST APIs, etc.
Version control of code and data, etc.
Hardware & resources
Scalability, GPU support, performance, etc.
Security & IT
On-premise, AD connectivity, cloud readiness, etc.
The recommended implementation concept is characterized by a hybrid stack of open source and enterprise solutions. This combines the required flexibility with good maintainability. The anchoring of a research environment for the uncomplicated and risk-free testing of new packages and functions by the data scientists should be emphasized.
The realization concept of eoda | analytic infrastructure consulting also includes a detailed plan for the implementation of the designed infrastructure and its components.
By professionalizing its IT infrastructure, AOK Baden-Württemberg is taking the framework conditions for data science to a new level. eoda’s implementation concept is the foundation for a high-performance, scalable and at the same time secure IT infrastructure that is perfectly designed for the productive use of data science in the company.
More about data infrastructures
Our offer: eoda | analytic infrastructure consulting
Concept, implementation and operation: We realize data infrastructures for the professional use of data science and AI in your company - performant, scalable, secure.
Case Study: Development of an analysis environment for group structures
We have built a powerful analysis environment for REWE Group around the productive use of the data science language R.
Case Study: Development of a centralized data infrastructure
For Covesto, we created a reproducible and centralized development environment for R and Python that helps address compliance guidelines while improving collaboration.
Blog: Background knowledge on data infrastructures and more
The data analysis blog from the data science specialist: Here you can learn more about data science in a business context, technological developments, the productive use of the data science languages R and Python, and more.