3rd Annual Data Science Day

Presenters: Davaughn Johnson

Faculty Advisor: Dr. Julia Shaedon

Abstract: Data Mining and Scarping are principles in Data Science that have created important tools. Some of these tools are user data scarping, extraction of data, and putting data into classes. In this presentation, we will go over how these tools have affected the world market and the potential impact the tools could have on companies' ethics.

Presenters: Jack Galloway

Faculty Advisor: Dr. Jason Polk, Department of Earth, Environmental, and Atmospheric Sciences

Abstract: This project focuses on the development of an advanced water resources monitoring dashboard, emphasizing the creation of an integrated, interactive platform for real-time data analysis and visualization. Situated in Bowling Green, Kentucky, where karst features present unique challenges, the dashboard serves as a vital tool for understanding and managing water resources. Leveraging NexSens/Onset Data Loggers and cloud-based platforms like WQDataLive and Hobolink, extensive data on water levels, rainfall, and water quality is collected. An API facilitates seamless integration of this data into the dashboard, which is further enhanced through Python scripting in ArcOnline for automated data analysis. Individual dashboards for each monitoring site are then integrated into a unified interface using ArcGIS Experience, ensuring accessibility and user engagement. Through this project, advancements in water resource monitoring and management are achieved, highlighting the pivotal role of the comprehensive dashboard in addressing critical challenges in karst areas. This innovative approach contributes to improved decision-making and community resilience in the face of water-related hazards.

Presenters: Nima Esmaeilzadeh

Faculty Advisor: Dr. Nahid Gani, Department of Earth, Environmental, and Atmospheric Sciences

Abstract: Geological mapping is crucial for mineral exploration, particularly in identifying lithological units, alterations, and critical minerals. Traditional methods are often limited by accessibility, cost, and environmental factors. This research demonstrates the effectiveness of leveraging multi-sensor remote sensing satellite imagery as a cost-efficient alternative for detailed mapping of mineral alteration zones. Focusing on the Mountain Pass District in California's eastern Mojave Desert, the research addresses a gap in comprehensive multi-satellite multispectral remote sensing analysis of the geologic framework related to carbonatite ore deposits and hydrothermal alteration zones. Using data from Landsat-9, Sentinel-2, and the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), the study successfully mapped iron-oxide and ferrous-bearing minerals, as well as hydrothermally altered rocks like clay minerals. Advanced machine learning algorithms such as Support Vector Machine and Random Forest significantly improved the classification and mapping of hydrothermal alteration minerals, demonstrating their effectiveness in geological exploration. As this research enhances our understanding of multispectral satellite data in mineral exploration, it establishes a robust framework for the future identification and characterization of mineral-rich zones. More broadly, the advancements heralded by this study are instrumental in reducing the United States' dependency on foreign mineral sources, thereby enhancing national security and supporting the sustainable development of crucial technologies and industries.

Presenters: Beth Rogers, Senior Manager of Data Science, Bath & Body Works

Abstract: Join Beth Rogers, head of Data Science at Bath & Body Works, as she takes us through her unique career journey. Learn what it's like to be a scientist in a top marine biology lab, brain imaging research center and cave research program. See how Beth translated her academia and scientific skillsets to a successful career as a data scientist and leader in corporate industry. Find out what types of business questions are being asked and techniques data scientists are using every day to answer them. Discover how you can start your own data science journey and begin your career path ideation.

Presenters: Shake Ibna Abir

Faculty Advisor: Dr. Richard Schugart

Abstract: Acute ischemic stroke, originating from cerebral artery blockage, often culminates in long-term disability or death. Identifying abnormal, or at-risk tissue in patients typically involves employing Computed Tomography Perfusion (CTP) imaging in patient-specific assessments. Perfusion images help identify areas of abnormality, including irreversibly-damaged tissue and tissue at risk, for acute brain stroke patients. The most significant physiologic and pathophysiologic estimated perfusion parameters are Cerebral Blood Volume (CBV), Cerebral Blood Flow (CBF), Time to Peak (TTP) and Mean Transit Time (MTT). The recent advancements in medical imaging have been propelled by the breakthroughs in Deep Learning (DL), particularly through the advent of Temporal Convolution Neural Networks (TCNN). This study presents an investigation into the calculated perfusion parameters extracted from Time Attenuation Curve (TAC) data for an initial patient using TCNN, Convolution Neural Networks (CNN), and Recurrent Neural Networks (RNN). A comparison of the methods will be presented as the corresponding root mean square errors (RMSE) for TCNN, CNN, and RNN are 0.05, 0.07, and 0.38, respectively.

Presenters: JC Watkins, Kentucky APEX Accelerator, Kentucky Science & Technology Corporation

Abstract: This non-research talk is designed to assist data scientists in considering what to do with the skills and abilities that have been developed while attending and working at WKU.  Graduation and post-graduation opportunities abound for data scientists. As a data scientist, you'll have a plethora of job opportunities across various industries such as tech, healthcare, finance, retail, etc. You could work as a data analyst, machine learning engineer, AI researcher, or data engineer, depending on your skills and interests. 

Presenters: Charlie Yielding, Data4All Podcaster

Abstract: Data will be trusted in the workforce of the future, but what about now? In this non research presentation, Charlie Yielding will share what it's like in the workforce in 2024 while giving examples through data storytelling.  At the end of this conversation, you will understand the dilemma you currently face, what the data says, and will leave with actions to take in order to prepare for your dream job.

Presenters: Luna Asbell, Katie Isaacs

Faculty Advisor: Dr. Jeremy Maddox, Department of Chemistry

Abstract: GaussianTools is a Mathematica package for processing data produced by the electronic structure software package Gaussian.  In this presentation we demonstrate an application of GaussianTools for visualizing the point group symmetry elements and operations of molecular structures.  We envision that instructors and students will be able to design interactive Mathematica notebooks to enhance their understanding of molecular symmetry.

Presenters: Cameron White, Dollar General Corp, WKU alum

Abstract: Often in academia, the size of datasets under study are “just right”: large enough to glean many insights from, yet small enough to be worked on with a single desktop PC. In this talk, experienced data scientist Cameron White will discuss various approaches on how to handle those situations when the datasets are not just right. At some firms, the datasets aren’t really that large and comprehensive, thus hampering some of the questions that can be asked. At others, there is too much data, which can be very difficult to wrangle on a single machine. 

Presenters: Cheyla Ferguson

Faculty Advisor: Dr. Amy Brausch, Department of Psychological Sciences

Abstract: Introduction/problem: Individuals with disordered eating behaviors face a higher risk of suicidal ideation, compared to the general population (Munick, 2023). Previous research has shown certain facets of eating disorders can predict suicidal ideation (Joiner et al., 2022). The Eating Disorder Inventory (EDI) includes multiple traits common in eating disorders such as low self-esteem, interpersonal alienation, and perfectionism. Considering the EDI incorporates questions about food but also about general thoughts and feelings, this study aims to uncover which subscale of the EDI is the best predictor of suicidal ideation. Based on previous research, we hypothesized that a subscale specific to eating disorders (i.e., interoceptive deficits, body dissatisfaction, or bulimia) will best predict suicidal ideation (Perkins & Brausch, 2019; Perkins et al., 2021; Rufino et al., 2018).