Abstract:

Test- and Measurement Data Problem:

Test- and measurement data is acquired during the development process of every product from simple toothbrushes up to complex machines and vehicles. Often, this data is captured in various formats such as XLSX, CSV, and MDF, making it challenging to add value, especially for further analysis and machine learning (ML).

Solution:

In this talk, we will present how ASAM ODS can help integrate different measurement data files into a holistic data view.  

Path to Solution:

Lightweight ETL pipelines based on microservices extract the existing data and enrich it with additional metadata during the import process. The imported data files can remain in their original location and format, preserving existing tool chains.

Once the data arrives in the ASAM ODS server, the existing HTTP(s)-REST-API provides secure data access from any location. However, in order to integrate this data with existing analysis and ML tools, another building block is needed to expose the data to the lingua franca of data scientists and data analysts – Python.

This gap is filled by the open-source ASAM ODSBox. This lightweight Python wrapper on top of the ASAM ODS HTTP-API introduces a simple JSON query language that makes it easy to learn and use ASAM ODS data. Additionally, the measurement data is returned as pandas.DataFrames, which naturally integrate with existing Python analytics and ML libraries.

This solution offers even more benefits. With support for the Python language, the data can now be easily used in interactive notebooks such as Jupyter or data analysis tools like Microsoft Power BI or MATLAB®.