Data Science and AI Services

The FDZ's Data Science Service provides expert evaluations of data science sections in scientific funding applications. Our team helps applicants optimize the quality and feasibility of their data strategy. This includes evaluating data collection methods, analytical approaches, visualization strategies, and technical infrastructure.

A data science pipeline involves several steps, from collecting and analyzing data to presenting results. The process begins with collecting and preparing the data, followed by analyzing and modeling it. Finally, the results are prepared for visual presentation and communication.

Summary of Services:

We support you in all phases of the data science pipeline, including data acquisition, cleansing, and analysis, as well as the deployment and training of AI systems. Additionally, we offer data visualization, publication consulting, and support for the FAIR sharing and archiving of your data in data repositories.

Reproducible and transparent research software and FAIR software development and research data. The FDZ supports software development by helping you write reproducible, open-source Python and R code and by organizing and documenting it in a structured way on GitHub. We also review your software repository for completeness and identify any issues. We support the FAIR publication and archiving of your research software and data in data repositories. Additionally, we can advise you on publishing an article about your research software and data in a scholarly journal, making your work accessible to a wider audience.

The FDZ offers consulting services for data science infrastructure, providing comprehensive support in selecting and using suitable computing resources, including high-performance computing (HPC). We advise on free cloud infrastructures, such as bwCloud, bwSync&Share, bwHPC, GWDG Cloud, Kaggle, Google Colab, Google Cloud for Education, and the TPU Research Cloud. Additionally, we provide expert guidance on efficiently using deep learning models in production environments to optimize and scale your projects.

Low-code and no-code approaches allow for the implementation of complex data science projects with minimal programming effort thanks to visual interfaces and pre-built libraries. Low-code tools require programming knowledge for customizations, while no-code platforms allow users without programming experience to create and use powerful data science models.

Our consulting services include support for selecting and applying low-code libraries, such as PyCaret, H2O AutoML, Auto-ViML, TPOT, and AutoKeras. We also advise on no-code tools, including Google Cloud Auto ML, ML KIT, Runway AI, Lobe, CreateML, RapidMiner, and DataRobot. Additionally, we offer consulting services on the efficient use of chatbots, such as ChatGPT and its alternatives, for research tasks and data analysis.

We offer comprehensive networking support in the field of data science. This includes helping you find suitable partners to promote your projects and leverage synergies. We also support you in organizing data science meetings to promote knowledge and experience exchange within the community and help you establish valuable contacts.

An additional data science partner at the University of Mannheim is the Mannheim Center for Data Science.