As a Post-Bachelor's Research Associate at Oak Ridge National Laboratory, I am helping to estimate populations worldwide by contributing to the Geographic Information Information and Sciences' LandScan HD product. In my focus country of Yemen, I researched relevant geographic data for population estimation in Yemen including land cover and census information and constructed high-resolution GIS borders for Yemen and its subsidiary administrative regions for precise population density estimations. I also generated over 20,000 high-resolution GIS polygon samples for the training of a building-detection deep learner in Yemen and North Korea via remote sensing.
Simultaneously to my digitizing, I learned python and was assigned the task of working collaboratively to generate automated workflows to generate population estimations for target countries worldwide. Through the use of Esri's ArcGIS spatial python package, I helped publish the internal python package lshdpopwork that would generate custom borders for a target country, glean a variety of land-use data that corresponds with those borders, and reformat that land-use data for downstream population estimations. As a component of both the border creation and land use layer preparation processes, I helped to construct a PostgreSQL data base and load a self-updating copy of OpenStreetMap.org data into it using PostGIS.
I was invited back to Oak Ridge in the summer of 2019 for a short-term contract to help transition the LandScan HD workflow away from using ArcGIS to using exclusively PostgreSQL and PostGIS. While much of the land-use workflows had already experienced this transition, I worked to move the border creation process into a python/SQL hybrid format. I recreated a PostgreSQL database to house the planet copy of OSM.org data and authored code to extract borders from there and other government-sanctioned spatial border datasets to generate worldwide and self-updating custom borders specifically for use within the LSHD workflow using PostGIS. I helped publish an updated version of the lshdpopwork python module that fully used PostGIS rather than ArcGIS for increased automation and quicker computational processing.
While not officially classified, no examples of geographic data are given here due to their non-reproducible status mandated by ORNL and the US Department of Energy. Research products are For Official Use Only (FOUO). Inquiries can be made to my mentor, Eric M. Weber, at [email protected] , for further details.
Simultaneously to my digitizing, I learned python and was assigned the task of working collaboratively to generate automated workflows to generate population estimations for target countries worldwide. Through the use of Esri's ArcGIS spatial python package, I helped publish the internal python package lshdpopwork that would generate custom borders for a target country, glean a variety of land-use data that corresponds with those borders, and reformat that land-use data for downstream population estimations. As a component of both the border creation and land use layer preparation processes, I helped to construct a PostgreSQL data base and load a self-updating copy of OpenStreetMap.org data into it using PostGIS.
I was invited back to Oak Ridge in the summer of 2019 for a short-term contract to help transition the LandScan HD workflow away from using ArcGIS to using exclusively PostgreSQL and PostGIS. While much of the land-use workflows had already experienced this transition, I worked to move the border creation process into a python/SQL hybrid format. I recreated a PostgreSQL database to house the planet copy of OSM.org data and authored code to extract borders from there and other government-sanctioned spatial border datasets to generate worldwide and self-updating custom borders specifically for use within the LSHD workflow using PostGIS. I helped publish an updated version of the lshdpopwork python module that fully used PostGIS rather than ArcGIS for increased automation and quicker computational processing.
While not officially classified, no examples of geographic data are given here due to their non-reproducible status mandated by ORNL and the US Department of Energy. Research products are For Official Use Only (FOUO). Inquiries can be made to my mentor, Eric M. Weber, at [email protected] , for further details.