Getting Fusion-Ready for Data-Intensive Geo-Spatiotemporal Analysis

Name: The ESSIC Seminar Series is on Winter Break!
Start: 2026-01-26
End: 2026-01-26

This event has passed. See the seminar recording here:

Dr. Kwo-Sen Kuo

NASA Goddard Space Flight Center

Monday October 7, 2019, 12 PM

ESSIC Conference Room 4102, 5825 University Research Ct, College Park, MD 20740

Abstract:

“Most data scientists spend only 20 percent of their time on actual data analysis and 80 percent of their time finding, cleaning, and reorganizing huge amount of data, which is an inefficient data strategy.”

– Armand Ruiz, The 80/20 Data Science Dilemma

The 80/20 dilemma may often be seen as a result of data volume, but it is primarily caused by data variety. Parallel processing is the only effective means to address the data volume challenge. However, without simultaneously addressing the data variety challenge, that effectiveness is significantly curtailed. In Earth Science, variety results from the different grids (including resolution) used by various models and from observations made by a large assortment of instruments on various platforms, e.g. ground-based, airborne, and spaceborne. Observation data need to be analyzed and assimilated to initialize model simulations and model outputs need to be analyzed with observations for verification and validation. Both observations and model outputs must be used in analyses to deepen understanding of and gain insights into the Earth system. Since Earth Science is a system science, a comprehensive approach demands integrative analyses of diverse data from multiple subdisciplines. The variety challenge is thus especially acute, and the 80/20 dilemma has become our daily experience.

I plan to conduct a demonstration in the first ~20 minutes of this seminar to show the key issue using “geo-lly” beans with audience participation. It will be followed with a PowerPoint presentation to introduce a recent innovation that can effectively “harmonize” the variety in geo-spatiotemporal data, thus rendering them fusion-ready to facilitate efficient data-intensive analysis.

https://www.infoworld.com/article/3228245/the-80-20-data-science-dilemma.html

Bio-sketch:

Kwo-Sen Kuo considers himself a “disruptive thinker” (commonly known as “boat-rocker” or “troublemaker”) because he likes to question the existing ways of doing things. Although he considers that to be completely rational, it is not always appreciated as so by others. His disruptiveness hasn’t mellowed with the advancement of age, which he himself is curious about and questioning why. He has been working with voluminous and diverse geo-data since he was exposed to LANDSAT and AVHRR data, considered BIG at the time too, when he was getting his master’s degree. He learned and applied then various data processing techniques that are now fashionably classified as Machine Learning and/or Artificial Intelligence. He got his PhD degree in Atmospheric Science from Colorado State University, specialized in Radiative Transfer. He is now affiliated with NASA Goddard Space Flight Center through University of Maryland-College Park. He is also the Chief Data Scientist and C*O of his company, Bayesics LLC (where “*” is a wildcard).

Webinar info:

Event site: http://go.umd.edu/kuo

Event number: 731 248 521
Event password: essic
——————————————————-
To join the online event
——————————————————-
1. Click here to join the online event.
Or copy and paste the following link to a browser:
https://umd.webex.com/umd/onstage/g.php?MTID=e2f4fa211a3bb9e89813ef9c6e6c428c0
2. Click “Join Now”.
——————————————————-
To join the audio conference only
——————————————————-
US Toll: +1-415-655-0002
Global call-in numbers:

https://umd.webex.com/umd/globalcallin.php?MTID=e28b990ce6a3825481a5e8e1a58471e95
Access code: 731 248 521
——————————————————-
For IT assistance
——————————————————-
Cazzy Medley: cazzy@umd.edu