79297 research outputs found
Sort by
Computational Data Analysis
No full textData science has emerged as a cornerstone of innovation, shaping an ever-expanding range
of professional careers. As technology advances and the volume of data expands expo-
nentially, the ability to extract meaningful insights from data has become indispensable
across industries. Far from representing a single career path, data science enables profes-
sionals in nearly every domain to make informed decisions, optimize systems, and drive
innovation. Yet, many high school students and incoming college freshmen have limited
exposure to data science fundamentals or the career opportunities they unlock. This is
the gap that Computational Data Analysis, a high school-level curriculum I developed,
seeks to address.
This curriculum originated during the 2022–2023 academic year, when I was awarded
a Public Service Assistantship through the Department of Mathematics at Louisiana
State University in partnership with the Gordon A. Cain Center. As part of this role, I
taught LSU STEM Pathway courses at Liberty High School in Baton Rouge, including
Data Manipulation and Analysis, a course originally designed by Mr. Alegre for the
LSU STEM Pathways program. Following his departure in 2023–2024, I redesigned and
expanded the curriculum to better prepare students for both professional and academic
pathways in data science, particularly LSU’s newly launched undergraduate data science
major.
Computational Data Analysis introduces students to core principles in data analysis,
computing, and statistical thinking, while incorporating industry-standard tools such as
R. The course culminates in the Certified Internet Web Professional (CIW) Data Analyst
certification exam, a nationally recognized credential that enhances students’ credibility in
both academic and professional contexts. Based on classroom experience, the curriculum
demonstrates strong potential for high certification pass rates and meaningful student
engagement.
Beyond the previously discussed certification benefits, the course also functions as a
bridge to collegiate-level degree paths, aligning with a nation-wide educational shifts to-
ward earlier specialization in data science. Whereas students once pursued general STEM
degrees before entering data-focused fields, new undergraduate majors now provide direct
pathways into the discipline. This curriculum serves as a practical precursor to those pro-
grams, which give students earlier exposure to the analytical and technical competencies
essential for success.
In conclusion, with Computational Data Analysis, I have sought to contribute to math-
ematics and statistics education by designing a curriculum that helps prepare high school
students to thrive in a data-driven world and to pursue further study in quantitative
disciplines. The long-term vision is to refine and distribute this curriculum through a
digital platform, making it accessible to high schools across Louisiana and, eventually the United States
Explainable AI-Enabled Forecasting of Dissolved Oxygen for Sustainable Aeration in a Rural Wastewater Treatment System
No full textAccurate forecasting of dissolved oxygen (DO) is crucial for maintaining biological treatment and minimizing energy consumption in wastewater systems. This challenge is acute in rural stabilization ponds, where DO responds nonlinearly to nutrient fluctuations, episodic inflows, and seasonal variation. To address this, an interpretable forecasting framework was developed using a transformer-based foundation model for time series, fine-tuned on nearly one year of multivariate sensor data from a rural wastewater facility. Four seasonal models (spring, summer, fall, and winter) were trained and evaluated at a 24-hour horizon, corresponding to daily operational planning, with ablation experiments conducted at 1 hour and 168 hours to assess robustness. Benchmark comparisons against Support Vector Regression (SVR), XGBoost, Long Short-Term Memory (LSTM), and the transformer-based Temporal Fusion Transformer (TFT) demonstrated substantial improvements. At 24 h, the symmetric mean absolute percentage error (SMAPE) fell from 38–46% for classical machine learning baselines (SVR, XGBoost) and 16–25% for deep learning benchmarks (LSTM, TFT) to below 7% with the proposed framework, representing significant gains in accuracy and stability. Interpretability was systematically integrated: SHapley Additive exPlanations (SHAP) attributions identified pH, conductivity, temperature, turbidity, and ammonium as regime-specific drivers, while sensitivity analyses enabled actionable “what-if” exploration. These findings highlight the importance of integrating foundation modeling, seasonal segmentation, and SHAP-based interpretability in improving forecasting for data-constrained rural wastewater systems. The work provides a framework for transparent and decision-aligned DO prediction, with potential relevance for proactive aeration planning and sustainable operation of decentralized wastewater treatment systems
Comparing the Consistency between Age Estimations Across Various Three-Dimensional Model Creation Methods in an Attempt to find a Practical Proxy for Physical Remains
No full textThe purpose of this study is to determine whether digital three-dimensional models can serve as a viable alternative to physical human remains to allow continued research on fragile physical specimens. As digital three-dimensional models have become increasingly prevalent in anthropological and archaeological research, it is essential to assess their limitations and capabilities, particularly for subjective assessments that rely heavily on visual interpretation (Remondino & El-Hakim, 2006). While previous studies have focused on the use of digital three-dimensional models for quantitative measurements, their effectiveness for subjective analyses remains unclear (Cooper et al., 2003; Sholts et al., 2010; Stull et al., 2014; Reynolds et al., 2017; Omari et al., 2021). To address this possible limitation, cranial suture-based age estimation was selected as the primary metric in this study. This method is inherently subjective, relying on visual assessment rather than quantifiable measurements, making it an ideal test case for evaluating whether digital three-dimensional models can support interpretive judgments, not just objective data collection. Thirty digital three-dimensional models were created from ten physical crania using three different modeling techniques: traditional photogrammetry (Agisoft Metashape), cellular photogrammetry (RealityScan), and structured light scanning (Artec 3D Space Spider). Five cranial sutures were selected for analysis: the coronal pterica, sagittal obelica, lambdoidal asterica, zygomaticomaxillary, and interpalatine. Student participants scored the sutures on the physical crania and then repeated the scoring process using the thirty digital three-dimensional models during two separate scoring sessions. Linear mixed effects models were used to analyze the results and account for repeated measures and inter-observer variability. The results demonstrated that models created with the Artec 3D Space Spider most closely approximated the scores from the physical specimens across four of the five sutures. Specifically, the results indicate that the RealityScan models provided the most similar responses for the zygomaticomaxillary suture. Agisoft Metashape and RealityScan produced moderately comparable results, but their accuracy varied depending on the anatomical location of the suture and the individual cranium. RealityScan models were more prone to surface distortion in areas that required post-scan merging, which occasionally compromised scoring accuracy. These findings suggest that digital three-dimensional models can effectively be used to evaluate subjective morphological traits, such as cranial suture closure. However, the quality and reliability of the model are highly dependent on the method used for model creation. Structured light scanning offers the most consistent visual accuracy, while photogrammetry-based methods are viable alternatives under constrained conditions. This study supports the inclusion of digital three-dimensional models in anthropological research and provides a framework for evaluating their effectiveness in analyzing subjectively scored skeletal features
Analog Unruh effect of inhomogeneous one-dimensional Dirac fermions
No full textWe study one-dimensional Dirac fermions in the presence of a spatially varying Dirac velocity v(x), that can form an approximate laboratory-based Rindler Hamiltonian describing an observer accelerating in Minkowski spacetime. A sudden switch from a spatially homogeneous velocity [v(x) constant] to a spatially varying velocity [v(x) inhomogeneous] leads to the phenomenon of particle creation, i.e., an analog Unruh effect. We study the dependence of the analog Unruh effect on the precise form of the velocity profile, finding that while the ideal Unruh effect occurs for v(x) ∝ |x|, a modified Unruh effect still occurs for more realistic velocity profiles that are linear for |x| smaller than a length scale λ and constant for |x| ≫ λ [such as v(x) ∝ tanh (|x|/λ)]. We show that the associated particle creation is localized to (Formula presented
MODELING BIOMASS DEWATERING BEHAVIOR DURING CONFINED MECHANICAL COMPRESSION
No full textThis thesis presents a detailed investigation into the modeling of biomass dewatering during confined mechanical compression. The experimental framework was carefully designed to replicate the industrial compaction and dewatering process of biomass while ensuring accurate acquisition of compression and fluid flow data. Emphasis is placed on the development of both empirical models, derived from experimental observations, and physics-based models grounded in poromechanical theory. The methodology includes a systematic description of the experimental procedures, followed by rigorous model calibration and validation. The results offer valuable insights into the mechanical behavior of biomass under compressive loading and lay a robust foundation for process optimization and future research in biomass processing
Effects of carbon substitution on magnetic properties and magnetocaloric effects in Mn65-xGa17C18+x compounds
No full textWe present an investigation involving the tuning of the magnetic, magnetocaloric, and room-temperature structural properties of Mn65-xGa17C18+x (0 ≤ x ≤ 4) compounds prepared using a high-energy ball milling (HEBM) technique. This study indicates that the crystal structure of all the compounds can be described as an anti-perovskite cubic structure with the Pm-3m space group and the crystal cell volume decreases with increasing carbon concentration. The system shows a first-order structural phase transition at a temperature T = TM between two cubic phases having different magnetic structures. The phases are characterized by antiferromagnetic (AFM) and ferromagnetic (FM) -like behavior at low (T \u3c TM) and high (TM ˃ T) temperature regions, respectively. A suppression of the AFM phase was observed with increasing C concentration. The temperature-induced first-order transitions (FOTs) were found to possess a small thermal hysteresis in the magnetization (∼ 2–3 K) in an applied magnetic field of H = 50 kOe. Magnetic entropy changes estimated from isothermal magnetization curves indicate that the largest value of the magnetic entropy change of |ΔSM| = 2.1 J kg-1 K−1 for x = 4 with ΔH = 50 kOe, with a relative cooling power (RCP) of ∼ 190 J kg−1.Thus high-energy ball milling (HEBM), a scalable technique, has been demonstrated as a viable method to synthesize magnetocaloric materials with substantial RCP values
Using machine learning for long-term calibration and validation of water quality ecosystem service models in data-scarce regions
No full textWater quality ecosystem service (ES) modeling tools help inform freshwater management across landscapes. However, the validity of such models depends on the availability of water quality data for validation and calibration, limiting their application in regions where monitoring is limited. This study presents a methodological framework that combines machine learning (ML) and spatial extrapolation to enhance ES modeling in data-scarce contexts (https://github.com/LSU-EPG/Puerto-Rico-ES-Project/tree/main/Data_Scarcity_Framework). Focusing on Puerto Rico, we leverage ML to reconstruct temporal gaps in nutrient trends. We then use ML to apply this reconstructed data to automate calibration and validation of nutrient retention ES models. We transfer validated parameters to unmonitored catchments using hydrogeological similarity. Our results demonstrate that ML preserves critical patterns in nutrient dynamics, while using calibrated parameters across hydrologically similar basins yields accurate predictions in ungauged watersheds. Our framework enhances ES model scalability, offering a tool to inform evidence-based water quality management in data-limited regions
A Stronger Kuroshio Intrusion Leads to Higher Chlorophyll a Concentration in the Northern South China Sea
No full textThe Kuroshio intrusion from the Luzon Strait significantly affects ecosystems in the South China Sea (SCS), especially during the Northeast Monsoon, a time when field observations are notably sparse and where vertical mixing induced by strong winds can obscure the effects of the Kuroshio intrusion. In this study, we address these gaps by reanalyzing data from 20 cruises (5,067 samples) in the SCS between 2004 and 2015. We also carried out two dedicated field cruises during the Northeast and the Southwest Monsoon in 2018. Field observations from both cruises revealed a consistent unimodal relationship between total chlorophyll a (Chla) concentrations in the upper 50 m of the water column and the index of the Kuroshio intrusion. Specifically, a strong Kuroshio intrusion during the Northeast Monsoon significantly enhanced Chla concentrations in the northern SCS. This enhanced Chla concentration during the Northeast Monsoon was primarily driven by increases of Synechococcus and nanophytoplankton that contrasted with the dominance of Prochlorococcus during the Southeast Monsoon. Long-term remote sensing data corroborated these findings and demonstrated a consistent pattern wherein intrusion by the Kuroshio led to elevated Chla concentrations, particularly during the Northeast Monsoon. There was a significant positive correlation between the intensity of the Kuroshio intrusion and the magnitude of the Chla increase. Furthermore, these findings suggested a concerning possibility: weakening of the Kuroshio intrusion intensity over time might diminish future biogeochemical effects on SCS ecosystems. Continued monitoring and research will be crucial to understanding and responding to these changes
