Leveraging the process mining technique to optimize data preparation time in a database used as an automated data...

Area of Science:

• Geophysics and Data Science
• Application of process mining in seismology
• Big data management in earth sciences

Background:

• Seismic centers like the U.S. Geological Survey (USGS) manage large volumes of earthquake waveform data.
• Traditional manual processing of seismic databases is time-consuming.
• Efficient management of big scientific datasets is crucial in modern seismology.

Purpose of the Study:

• To develop and implement a seismic database using process mining techniques.
• To analyze and optimize the workflow for preparing, delivering, and processing earthquake waveform records.
• To reduce the time required for seismic database creation and management.

Main Methods:

• Utilized process mining techniques to analyze the workflow of seismic database creation.
• Collected and processed nearly 900 earthquake waveform records from the USGS (1999-2018).
• Developed custom software scripts to automate time-consuming database creation phases.

Main Results:

• Process mining reduced database creation time from 25 days (manual) to approximately 8 days.
• Automated software scripts streamlined data request, collection, preparation, and delivery.
• Successfully managed a big data volume of seismic records over a large geographical area.

Conclusions:

• Process mining is an effective methodology for optimizing seismic database workflows.
• The proposed approach significantly enhances efficiency in creating, storing, and delivering seismic data.
• This method offers a valuable solution for seismological and other data centers managing large scientific datasets.