Menu
Thor Osborn
Thor Osborn

Thor Osborn, CAP

Principal Systems Research Analyst
Sandia National Laboratories
Bio

He focuses his data science efforts on analysis and operations research to inform executive leadership on key business decisions.  Most of his recent work has been in workforce behavioral studies, modeling and analysis, compensation, and document management.  He holds a doctorate in biomedical engineering from the University of Washington and an MBA from the University of New Mexico, and is a Certified Analytics Professional.  He has been awarded three US patents and has published on diverse topics including biological sensing, MEMS actuators, seeker systems, analysis of terrorism, and workforce analysis.


The Voice of the Universe: Why more data may not help you

The education and practice of analytics is founded primarily on system models comprising stationary processes, yet virtually all real-world analytics challenges address Complex Adaptive Systems (CAS) that are by definition non-stationary. CAS and their pink noise footprints are fundamental to human organizations. Pink noise signals cannot generally be predicted based on short-term data and are autocorrelated, yielding drift and model failure. Excepting narrowly defined circumstances, pink noise signals may neither be blocked, quelled, nor fully anticipated, often leading to issues that may not be prevented, only managed.  Meta-level awareness of the CAS perspective can be a valuable tool for practitioners throughout the analytics life cycle. The analyst’s responsibilities and tasks within the workflow have been outlined in detail by the INFORMS Certified Analytics Practitioner (CAP) Program as the basis for certification testing. Examination of this analytics workflow through the CAS lens expands the scope of relevant considerations in project development and execution and lends support to recommended practices for sustainable analytics in the real Universe.  For example, mindfulness of the inherent optimism of stationary system assumptions, acceptance that some system behaviors are beyond prior human knowledge or control, selection of modeling methods and variables with robustness in mind, and building in adequate stability by treating the design of model maintenance as an essential priority.

Back To Top

Institute for Operations Research and the Management Sciences

5521 Research Park Drive, Suite 200
Catonsville, MD 21228 USA

phone 1 443-757-3500

phone 2 800-4INFORMS (800-446-3676)

fax 443-757-3515

email

Get the Latest Updates

Copyright INFORMS. All Rights Reserved