For the course Data Science in Practice at the Leiden Institute of Advanced Computer Science we developed a data pipeline to incorporate Patient Generated Health Data (PGHD) for patients with long term health problems in Nigeria with the existing VODAN-Africa architecture. This work was performed in cooperation with the ADAPT center and the VODAN-Africa project. The main result is the functioning pipeline prototype which is outlined in the figure presented below. The full work has been accepted for publication at the HEALTHINF 2024 conference in Italy and will be shared soon
This work was performed as my Second Master Project for the Master Astronomy & Data Science in combination with the European Space Agency. Here we developed a methodology based on a Gaussian Mixture Model to attempt to improve upon the previous best Hipparcos - Gaia DR3 crossmatch by Marrese et al. (2019).The resulting submitted thesis can be found
here
.
Simple Summary: With the increasing amount of data coming in from telescopes both ground- and space-based, it has become impossible for astronomers to perform by-hand in depth analyses of all of these observations. Nevertheless, the regions surrounding massive stars are immensely complex and are dominated by a multitude of stellar feedback forces at any point in time. Understanding these, and thereby improving our understanding of how heavy stars impact further star formation, requires the in-depth analyses. In this work we have attempted to partially automate this analysis by creating a model based on GMMis by Melichor & Goulding (link) that is able to identify similar pixels within such a region.
This work was performed as my First Master Project for the Master Astronomy & Data Science, the resulting thesis can be found here. It has since been expanded into an article that has been published by the Astrophysical Journal which can be accessed at doi.org/10.3847/1538-4357/ad003c. The corresponding code is findable on GitHub.
created with
WordPress Website Builder .