Öffentlicher Dienst

KIT 3DMM2O – Automated Data Clusters

Algorithm for structure and cluster detection in heterogeneous research data.

Client: Karlsruher Institut für Technologie (KIT) Team size: 1.000-5.000 Mitarbeiter Services: Data Engineering, KI-Integrationen, Forschung & Entwicklung

Reference Open website

Demo View demo

Tech Stack

Python Pandas NumPy Custom Clustering Data Engineering

Services

Data Engineering KI-Integrationen Forschung & Entwicklung

Overview

For the 3DMM2O Excellence Cluster, I developed an algorithm that analyzes, structures, and clusters research data without a prior schema. Results were presented at the E-Science Days Heidelberg. The system processes metadata from Zenodo repositories and identifies patterns across diverse data types.

Challenge

Research data from multiple groups comes in heterogeneous formats. Manual categorization is time-consuming and inconsistent.

Solution

Automated type detection and clustering algorithms process thousands of datasets, identifying patterns and groupings without predefined schemas.

Results & Impact

Over 10,000 datasets analyzed and clustered automatically. Research groups can now discover related work across the cluster efficiently.

More Projects

More cases from similar industries or with comparable technology.

To project overview

Agentino AI

AI Chat Assistant + CRM for trades businesses. Automates lead capture, pre-qualification and appointment scheduling. Runs as webchat, multi-tenant, GDPR-compliant.