IL3.1A - Automatic EDA System for Heat Pump Data



A functional automated Exploratory Data Analysis system for Bosch heat-pump manufacturing data. It integrates AutoML, local LLM-based metadata enrichment, an ML Readiness metric and FAIR data-maturity reporting, and was tested with real Nexeed MES datasets from Bosch.

  • Main Gain(s): Automated exploratory analysis, Improved FAIR data maturity, ML readiness assessment, Persistent reporting for industrial datasets

  • Core Contributions: (concept) E. Rocha; (design) V. Costa; (implementation & testing) V. Costa

Use Cases:

  • Real Nexeed MES datasets from Bosch Termotecnologia heat-pump production processes

    • Start TRL: 4 - Use of open source applications

    • Final TRL: 7 - Functional and tested on Bosch data

Main Features

  • Automated exploratory and descriptive-statistics reports for high-volume batch data
  • Local Large Language Models for dataset descriptions and metadata enrichment
  • AutoML functionality to assess predictive potential in industrial datasets
  • ML Readiness metric for objective data-science readiness assessment
  • Dataverse-based FAIR framework with persistent reporting
  • Integration into the Project 3 Control Tower

Images and Videos

Product Demonstration
Short video on the Public Demonstration