Erich Seamon
Digital Portfolio
Hello! My name is Erich Seamon. Im a Research Scientist @ the University of Idaho. I'd like to provide you with a few visualization examples of my research, which focuses on spatiotemporal modeling and machine learning, and its application to a number of overlapping areas, including climate, agriculture, human health.
My research lab site can be found here:
https://haclab.uidaho.edu
Project 1: PhD Research: Spatiotemporal Relationships of Climate to Agricultural Insurance Loss
Using over 2.8 million agricultural insurance records, this research analyzed spatiotemporal variations of insurance loss, and how climate can predict values.
Here we animate insurance loss for wheat insurance loss due to drought from 1989 to 2015 - to visually see the spatial variability.
Two publications outline this work, which is noted in my CV.
Project 1: PhD Research: Spatiotemporal Relationships of Climate to Agricultural Insurance Loss
Using over 2.8 million agricultural insurance records, this research analyzed spatiotemporal variations of insurance loss, and how climate can predict values.
Here we used a novel algorithmic method to select the most influential time periods to best predict wheat losses, for each county. Our models had over .90 R 2 accuracy.
Two publications outline this work, which is noted in my CV.
Project 2: Spatiotemporal modeling of health outcomes in Idaho: COVID-19
Part of a multi-year project funded by the State of Idaho, we have developed spatiotemporal models that estimate over 40 health factors at a county level - related to COVID-19 and tobacco usage. We've additionally constructed dynamic visualizations and an API for data access, with an associated publication in Population, Space and Place ( https://doi.org/10.1002/psp.2647). Dynamic dashboards, API access, and other info can be found at:
Project 2: Spatiotemporal modeling of health outcomes in Idaho: TOBACCO
Part of a multi-year project funded by the State of Idaho, we have developed spatiotemporal models that estimate over 40 health factors at a county level - related to COVID-19 and tobacco usage. We've additionally constructed dynamic visualizations and an API for data access, with an associated publication in Population, Space and Place ( https://doi.org/10.1002/psp.2647). Dynamic dashboards, API access, and other info can be found at:
Project 3: Pandemic Modeling and COVID-19
As part of our University of Idaho pandemic modeling team, we developed an interactive web site and dashboards to assess covid risk for the United States. Data was automatically pulled from differing sources, with animated maps, graphs, and outputs generated daily.
Three papers were generated from this work, noted in my CV.
Project 4: Novel Spatiotemporal COVID-19 Modeling
We developed a spatially weighted random forest methodology, to use 15 core variables (socioeconomics, political ideology, vaccine uptake, health conditions), to predict covid deaths across the US - and to assess how predictability spatially varies.
In this example from our paper (in review at Scientific Reports), we show modeled deaths during the Delta wave window.
Project 5: Risk Perception and COVID-19 using Structural Equation Modeling
In this research, we collected survey data from three states (Idaho, Texas, and Vermont), and constructed differing structural equation models (SEM) to explore how risk perceptions, political ideology, and rurality spatially varies related to COVID-19 deaths and cases. We have a publication of this work in PLOS One:
Project 6: Geospatial modeling of crime and social factors in relationship to obesity
Using longitudinal temperament data collected on children in the Seattle area, we looked at spatial variations of cortisol and the relationships to crime, economics, and other demographic factors. This work is published in the Journal of Pediatric Psychology:
Project 7: Climatic Factors and Spatial Relationships to Aphid Populations in the PNW
Using aphid data collected over the course of 4 years, we analyzed spatial changes of abundance and sizing, and how climate and landscape factors (cropping systems) are related. This work is published in the Journal of Economic Entomology
Project 8: Student Mentoring: Geospatial Modeling Workshop (BCB 503)
As part of a graduate level course (BCB 503), geospatial modeling approaches were presented, with code provided – to guide students in the use of differing spatial statistics techniques and visualizations. All code and materials can be found here:
