Eric
Gong
Developing Web Application Stacks for Hosting R-based Statistical Models
Abstract profile. Full document pending author claim.
Authors:
Eric Gong, Giovanni Parmigiani, Danielle Braun
Date Created:
2025-01-01
Course Title:
Professor:
Not specified
About Paper:
The increased ease of collecting vast quantities of patient-specificophisticated or aesthetic user interfaces. Hence the development data enables biostatisticians to develop highly accurate statisticalf alternative web application software stacks is critical. Open- models capable of predicting a patient’s predisposition and risk source R packages, such as Plumber, offer a promising alternative; for disease, often long before symptoms manifest. Such models Plumber can be used to create API endpoints that can interface are enabled by powerful statistical programming languages, such with R-based statistical models. When deployed on publicly as the R programming language. However, deployment of accessible servers, these API endpoints can then be queried by these models is difficult. Clinicians and patients often lack the moreweb-applicationorientedlanguagesandlibrariestraditionally technical knowledge and computational infrastructure necessary used in web development, such as React. By combining insights to deploy local copies of these models. RShiny Apps exist as from previous user-experience studies with software engineering an alternative, enabling statisticians who developed the models a best practices, we propose a software stack uniquely suited way to provide a web application interface to their models, which for deploying R-based statistical tools: bolstering computational clinicians and patients can then access through a web browser. efficiency, facilitating interface customizability, and ultimately However, RShiny is plagued with hardware limitations, imposed paving the path to a more user-friendly clinician and patient quotas on application up-time, and a limited capacity to create experience overall. Simulating Quantum Phenomena in Classical Systems: Flatband Formation in Kagome Twisted Bilayers Moana Gorezi, Philip Shenk, Jennifer Hoffman Harvard College | Currier House | Mechanical Engineering | 2028 Kagome twisted bilayers (KTBs) offer a promising platform dispersion in metamaterial structures closely mimics the quantum for exploring quantum phenomena, particularly the formation band structures of electrons in kagome lattices. This allows us to of flatbands at specific “magic angles” of interlayer rotation. computationally probe flatband conditions using classical systems. These flatbands are valuable for studying correlated electronic Our simulations reveal the emergence of a flatband when the twist behaviors such as superconductivity. To ease the fabrication angleisapproximately3.14degrees. Identifyingsuchhighermagic process, our goal is to identify relatively high magic angles, angles in KTBs holds promise for advancing room-temperature which can improve experimental efficiency. In this work, we superconducting materials. This, in turn, could support the use COMSOL Multiphysics to model the geometry of the KTB development of scalable and practical quantum technologies. Our supercell (two kagome layers stacked with a designated twist study contributes a computational framework for identifying twist angle) and simulate wave propagation using pressure acoustics in geometries likely to host novel quantum phases, pushing forward the frequency domain. Prior research shows that acoustic wave the search for realizable flatband systems.
Abstract:
The increased ease of collecting vast quantities of patient-specificophisticated or aesthetic user interfaces. Hence the development data enables biostatisticians to develop highly accurate statisticalf alternative web application software stacks is critical. Open- models capable of predicting a patient’s predisposition and risk source R packages, such as Plumber, offer a promising alternative; for disease, often long before symptoms manifest. Such models Plumber can be used to create API endpoints that can interface are enabled by powerful statistical programming languages, such with R-based statistical models. When deployed on publicly as the R programming language. However, deployment of accessible servers, these API endpoints can then be queried by these models is difficult. Clinicians and patients often lack the moreweb-applicationorientedlanguagesandlibrariestraditionally technical knowledge and computational infrastructure necessary used in web development, such as React. By combining insights to deploy local copies of these models. RShiny Apps exist as from previous user-experience studies with software engineering an alternative, enabling statisticians who developed the models a best practices, we propose a software stack uniquely suited way to provide a web application interface to their models, which for deploying R-based statistical tools: bolstering computational clinicians and patients can then access through a web browser. efficiency, facilitating interface customizability, and ultimately However, RShiny is plagued with hardware limitations, imposed paving the path to a more user-friendly clinician and patient quotas on application up-time, and a limited capacity to create experience overall. Simulating Quantum Phenomena in Classical Systems: Flatband Formation in Kagome Twisted Bilayers Moana Gorezi, Philip Shenk, Jennifer Hoffman Harvard College | Currier House | Mechanical Engineering | 2028 Kagome twisted bilayers (KTBs) offer a promising platform dispersion in metamaterial structures closely mimics the quantum for exploring quantum phenomena, particularly the formation band structures of electrons in kagome lattices. This allows us to of flatbands at specific “magic angles” of interlayer rotation. computationally probe flatband conditions using classical systems. These flatbands are valuable for studying correlated electronic Our simulations reveal the emergence of a flatband when the twist behaviors such as superconductivity. To ease the fabrication angleisapproximately3.14degrees. Identifyingsuchhighermagic process, our goal is to identify relatively high magic angles, angles in KTBs holds promise for advancing room-temperature which can improve experimental efficiency. In this work, we superconducting materials. This, in turn, could support the use COMSOL Multiphysics to model the geometry of the KTB development of scalable and practical quantum technologies. Our supercell (two kagome layers stacked with a designated twist study contributes a computational framework for identifying twist angle) and simulate wave propagation using pressure acoustics in geometries likely to host novel quantum phases, pushing forward the frequency domain. Prior research shows that acoustic wave the search for realizable flatband systems.
Source:
Harvard / Y260X/A447T / 2025
Topics:
model, flatband, web, patient, quantum, application, statistical, interface, kagome, angle, stack, user
