Data Science Practice Research Articles

Philosophy within Data Science Ethics Courses

There is wide agreement that ethical considerations are a valuable aspect of a data science curriculum, and to that end, many data science programs offer courses in data science ethics. There are not always, however, explicit connections between data science ethics and the centuries-old work on ethics within the discipline of philosophy. Here, we present a framework for bringing together key data science practices with ethics topics. The ethics topics were collated from 16 data science ethics courses with public-facing syllabi and reading lists. We encourage individuals who are teaching data science ethics to engage with the philosophical literature and its connection to current data science practices, which are rife with potentially morally charged decision points.

Bias and Fairness in AI Models: How can Bias in AI Models be Identified, Mitigated, and Prevented in Data Science Practices?

Artificial intelligence (AI) and machine learning (ML) systems are progressively used in different areas, going with basic choices that influence individuals' lives. In any case, these frameworks can sustain and try and fuel existing social predispositions, prompting uncalled for results. This paper looks at the wellsprings of predisposition in simulated intelligence models, assesses current methods for distinguishing and relieving inclination, and proposes an extensive structure for creating more pleasant simulated intelligence frameworks. By coordinating specialized, moral, and functional points of view, this exploration plans to add to a more evenhanded utilization of computer-based intelligence across various areas, guaranteeing that artificial intelligence driven choices are fair, straightforward, and socially dependable.

The role of artificial intelligence and data science in nanoparticles development: a review.

Artificial intelligence has revolutionized many sectors with unparalleled predictive capabilities supported by machine learning (ML). So far, this tool has not been able to provide the same level of development in pharmaceutical nanotechnology. This review discusses the current data science methodologies related to polymeric drug-loaded nanoparticle production from an innovative multidisciplinary perspective while considering the strictest data science practices. Several methodological and data interpretation flaws were identified by analyzing the few qualified ML studies. Most issues lie in following appropriate analysis steps, such as cross-validation, balancing data, or testing alternative models. Thus, better-planned studies following the recommended data science analysis steps along with adequate numbers of experiments would change the current landscape, allowing the exploration of the full potential of ML.

Harnessing the Power of Artificial Intelligence in Climate Change Mitigation: Opportunities and Challenges for Public Health

Artificial intelligence (AI) has emerged as a powerful tool for addressing the challenges posed by climate change and their impact on public health. By leveraging its capacity to analyze and predict climatic patterns, AI offers opportunities to enhance resource management and develop effective strategies for climate change mitigation. Moreover, AI can contribute to generating sustainable solutions that address the complex and interconnected nature of climate change. For example, AI can enable the optimization of energy consumption and facilitate the integration of renewable energy sources into existing systems. It can also support the development of climate models that provide timely and accurate predictions, enabling policymakers to implement proactive measures for disaster preparedness and response. Furthermore, AI-powered disease surveillance and mitigation techniques can improve public health outcomes by identifying patterns and trends in the spread of diseases in relation to climatic factors. However, the widespread adoption of AI-based solutions is not without challenges. Ethical concerns surrounding privacy and data ownership must be addressed, as the use of AI requires access to large datasets, raising potential privacy risks. Technical constraints, such as limited computational power and the need for sophisticated algorithms, also pose obstacles to the implementation of AI in climate change mitigation strategies. Furthermore, issues related to the accessibility and affordability of AI technologies must be resolved to ensure equitable distribution and maximize its potential impact on public health. To fully harness the power of AI in addressing climate change and improving public health outcomes, it is crucial to promote innovation, multidisciplinary collaboration, and open data science. Innovation can drive the development of new AI algorithms and technologies specifically tailored to address climate change challenges. Multidisciplinary approaches that bring together experts from diverse fields, including climate science, public health, and computer science, can foster a holistic understanding of complex systems and enable the design of comprehensive solutions. Finally, open data science practices, such as sharing data and algorithms, can facilitate collaboration and accelerate progress in mitigating climate change and its public health impacts. In conclusion, AI offers promising opportunities for effectively addressing climate change and its impact on public health. However, to fully realize its potential, it is essential to tackle ethical and privacy concerns, overcome technical constraints, and ensure accessibility and affordability. By promoting innovation, multidisciplinary collaboration, and open data science, we can unlock the transformative power of AI in mitigating climate change and improving public health outcomes.

102 CUBE: A Collaborative Undergraduate Biostatistics Experience to Bring Diversity and Awareness to the Field of Collaborative Biostatistics

OBJECTIVES/GOALS: Despite a steady rise of graduate degrees in biostatistics earned in the US, the percent from minorities remains low. This poster will describe the Collaborative Undergraduate Biostatistics Experience (CUBE), an 8-week program aimed to diversify and bring awareness to the field of collaborative biostatistics, from recruitment through evaluation. METHODS/STUDY POPULATION: The CUBE program is funded jointly by the NIH’s NIDA/NIAAA (award number: 1R25DA058482-01) and is designed to give underrepresented minority (URM) undergraduate students in STEM the opportunity to engage in a collaborative biostatistics and health data science experience, along with related professional development activities. The program is built on four pillars: 1) training in introductory biostatistics, 2) training in R programming, 3) professional development, and 4) a collaborative research project addressing research questions in various disciplines. The CUBE program was delivered in the summer of 2022 as a pilot to four URM students at Virginia Tech (VT) and the University of Virginia (UVA), with two at each site. In summer 2023, the program was offered to 5 students (3 VT, 2 UVA). RESULTS/ANTICIPATED RESULTS: This poster will provide strategies learned over two summers with respect to recruitment and enrollment, along with details on the program content, timeline, and short- and long-term program evaluation metrics (both quantitative and qualitative). The CUBE program was well-received by students participating in summers 2022 and 2023, where improved attitudes towards statistics were demonstrated, and 7 of the total 9 participants (78%) over the past two summers expressed interest in pursuing a graduate degree in biostatistics or a career in quantitative research. Of these 7 students, 1 is currently enrolled in a biostatistics graduate program in the United States. DISCUSSION/SIGNIFICANCE: Results can be used to offer recommendations to leaders in the field on how to establish similar programs seeking to provide a pipeline for equity and diversity in the practice of collaborative biostatistics and health data science.

Abstract 1615: Balanced molecular feature selection unveils universally tumor-lineage informative methylation sites in colorectal cancer

Abstract In the era of precision medicine, performing comparative analysis over diverse patient populations is a fundamental step toward tailoring healthcare interventions. However, the critical aspect of equitably selecting molecular features across multiple patients is often overlooked. To address this challenge, we introduce FALAFL (FAir muLti-sAmple Feature seLection), a novel algorithmic approach based on integer linear programming. FALAFL is designed to bridge the gap between molecular feature selection and algorithmic fairness, ensuring a balanced selection of molecular features across different patient samples. We have applied FALAFL to the problem of selecting lineage-informative CpG sites within a cohort of 9 colorectal cancer patients subjected to low-coverage single-cell methylation sequencing. Our results demonstrate that FALAFL excels at rapidly determining the optimal set of CpG sites, which are well covered by cells across the vast majority of the patients, while ensuring that each patient contributes a similar number of sites to the final selection. An analysis of the FALAFL-selected CpG sites reveals that their lineage informativeness exhibits a strong association across diverse patient profiles, suggesting the non-stochasticity and universality of methylation changes in tumor evolution. Furthermore, these universally lineage-informative sites exhibit subclonal level alterations in the colorectal cancer primary tumors and are strongly associated with lineage-specific gene expression patterns. By integrating equity considerations into the molecular feature selection process, FALAFL is poised to propel equitable healthcare data science practices and advance our understanding of cancer and other complex diseases. Citation Format: Xuan Cindy Li, Yuelin Liu, Alejandro A. Schäffer, Stephen M. Mount, Süleyman C. Sahinalp. Balanced molecular feature selection unveils universally tumor-lineage informative methylation sites in colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 1615.

Fundamental Components and Principles of Supervised Machine Learning Workflows with Numerical and Categorical Data

This paper offers a comprehensive examination of the process involved in developing and automating supervised end-to-end machine learning workflows for forecasting and classification purposes. It offers a complete overview of the components (i.e., feature engineering and model selection), principles (i.e., bias–variance decomposition, model complexity, overfitting, model sensitivity to feature assumptions and scaling, and output interpretability), models (i.e., neural networks and regression models), methods (i.e., cross-validation and data augmentation), metrics (i.e., Mean Squared Error and F1-score) and tools that rule most supervised learning applications with numerical and categorical data, as well as their integration, automation, and deployment. The end goal and contribution of this paper is the education and guidance of the non-AI expert academic community regarding complete and rigorous machine learning workflows and data science practices, from problem scoping to design and state-of-the-art automation tools, including basic principles and reasoning in the choice of methods. The paper delves into the critical stages of supervised machine learning workflow development, many of which are often omitted by researchers, and covers foundational concepts essential for understanding and optimizing a functional machine learning workflow, thereby offering a holistic view of task-specific application development for applied researchers who are non-AI experts. This paper may be of significant value to academic researchers developing and prototyping machine learning workflows for their own research or as customer-tailored solutions for government and industry partners.

Methodology for Good Machine Learning with Multi-Omics Data.

In 2020, Novartis Pharmaceuticals Corporation and the U.S. Food and Drug Administration (FDA) started a 4-year scientific collaboration to approach complex new data modalities and advanced analytics. The scientific question was to find novel radio-genomics-based prognostic and predictive factors for HR+/HER- metastatic breast cancer under a Research Collaboration Agreement. This collaboration has been providing valuable insights to help successfully implement future scientific projects, particularly using artificial intelligence and machine learning. This tutorial aims to provide tangible guidelines for a multi-omics project that includes multidisciplinary expert teams, spanning across different institutions. We cover key ideas, such as "maintaining effective communication" and "following good data science practices," followed by the four steps of exploratory projects, namely (1) plan, (2) design, (3) develop, and (4) disseminate. We break each step into smaller concepts with strategies for implementation and provide illustrations from our collaboration to further give the readers actionable guidance.

Engaging elementary students in data science practices

PurposeThis study is part of a participatory design research project and aims to develop and study pedagogical frameworks and tools for integrating computational thinking (CT) concepts and data science practices into elementary school classrooms.Design/methodology/approachThis paper describes a pedagogical approach that uses a data science framework the research team developed to assist teachers in providing data science instruction to elementary-aged students. Using phenomenological case study methodology, the authors use classroom observations, student focus groups, video recordings and artifacts to detail ways learners engage in data science practices and understand how they perceive their engagement during activities and learning.FindingsFindings suggest student engagement in data science is enhanced when data problems are contextualized and connected to students’ lived experiences; data analysis and data-based decision-making is practiced in multiple ways; and students are given choices to communicate patterns, interpret graphs and tell data stories. The authors note challenges students experienced with data practices including conflict between inconsistencies in data patterns and lived experiences and focusing on data visualization appearances versus relationships between variables.Originality/valueData science instruction in elementary schools is an understudied, emerging and important area of data science education. Most elementary schools offer limited data science instruction; few elementary schools offer data science curriculum with embedded CT practices integrated across disciplines. This research assists elementary educators in fostering children's data science engagement and agency while developing their ability to reason, visualize and make decisions with data.

Data Flourishing: Developing Human‐Centered Data Science through Communities of Ethical Practice

ABSTRACTData Science is a burgeoning area in the iField. But Data Science practices have far outstripped the field's ethical safeguards. We argue that Data Science graduate education programs must address this critical problem. In this theoretical and conceptual paper, we posit an ordinary macroethics that we call data flourishing. We contend that this macroethics is most appropriately developed through a holistic, human‐centered data science (HCDS)‐based pedagogy that concentrates on cultivating communities of ethical practice (COEPs) through social learning. We favor embedding this macroethics throughout iField programs' graduate data science curricula and by extension, the entire data science education enterprise. This paper aligns with the 2023 ASIS&T annual meeting theme of translating research into practice, particularly the subthemes of “improving decision‐making” and “understanding the power of information to develop human happiness, equality, and wellbeing.”

What is uncertainty in today’s practice of data science?

What is uncertainty in today’s practice of data science?

Open Science Best Practices in Data Science and Artificial Intelligence

In the past years, scientific research in Data Science and Artificial Intelligence has witnessed vast progress. The number of published papers and digital objects (data, code, models) is growing exponentially. However, not all of these research artifacts are findable, accessible, interoperable and reusable (FAIR), contributing to a rather low level of reproducibility of the experimental findings reported in scholarly publications (reproducibility crisis). In this paper, we focus on Data Science and Artificial Intelligence Open Science best practices, i.e., a set of recommendations that eventually contribute to the management and development of digital artefacts that are as FAIR as possible. While several guidelines exist, we add best practices for the FAIR collection, processing, storing and sharing of scholarly findings via Research Knowledge Graphs. The final list of recommendations will be openly available on the NFDI4DS webpage as an interactive web application.

Data science pedagogical tools and practices: A systematic literature review

Data science pedagogical tools and practices: A systematic literature review

Analytics‐supported reflective assessment for 6th graders' knowledge building and data science practices: An exploratory study

AbstractPreparing data‐literate citizens and supporting future generations to effectively work with data is challenging. Engaging students in Knowledge Building (KB) may be a promising way to respond to this challenge because it requires students to reflect on and direct their inquiry with the support of data. Informed by previous studies, this research explored how an analytics‐supported reflective assessment (AsRA)‐enhanced KB design influenced 6th graders' KB and data science practices in a science education setting. One intact class with 56 students participated in this study. The analysis of students' Knowledge Forum discourse showed the positive influences of the AsRA‐enhanced KB design on students' development of KB and data science practices. Further analysis of different‐performing groups revealed that the AsRA‐enhanced KB design was accessible to all performing groups. These findings have important implications for teachers and researchers who aim to develop students' KB and data science practices, and general high‐level collaborative inquiry skills. Practitioner notesWhat is already known about this topic Data use becomes increasingly important in the K‐12 educational context. Little is known about how to scaffold students to develop data science practices. Knowledge Building (KB) and learning analytics‐supported reflective assessment (AsRA) show premises in developing these practices. What this paper adds AsRA‐enhanced KB can help students improve KB and data science practices over time. AsRA‐enhanced KB design benefits students of different‐performing groups. AsRA‐enhanced KB is accessible to elementary school students in science education. Implications for practice and/or policy Developing a collaborative and reflective culture helps students engage in collaborative inquiry. Pedagogical approaches and analytic tools can be developed to support students' data‐driven decision‐making in inquiry learning.

New Vistas on Responsible Learning Analytics

The focus of ethics in learning analytics (LA) frameworks and guidelines is predominantly on procedural elements of data management and accountability. Another, less represented focus is on the duty to act and LA as a moral practice. Data feminism as a critical theoretical approach to data science practices may offer LA research and practitioners a valuable lens through which to consider LA as a moral practice. This paper examines what data feminism can offer the LA community. It identifies critical questions for further developing and enabling a responsible stance in LA research and practice taking one particular case — algorithmic decision-making — as a point of departure.

Three Principles for Modernizing an Undergraduate Regression Analysis Course

As data have become more prevalent in academia, industry, and daily life, it is imperative that undergraduate students are equipped with the skills needed to analyze data in the modern environment. In recent years there has been a lot of work innovating introductory statistics courses and developing introductory data science courses; however, there has been less work beyond the first course. This article describes innovations to Regression Analysis taught at Duke University, a course focused on application that serves a diverse undergraduate student population of statistics and data science majors along with nonmajors. Three principles guiding the modernization of the course are presented with details about how these principles align with the necessary skills of practice outlined in recent statistics and data science curriculum guidelines. The article includes pedagogical strategies, motivated by the innovations in introductory courses, that make it feasible to implement skills for the practice of modern statistics and data science alongside fundamental statistical concepts. The article concludes with the impact of these changes, challenges, and next steps for the course. Portions of in-class activities and assignments are included in the article, with full sample assignments and resources for finding data in the supplemental materials. Supplementary materials for this article are available online.

Good Data Science Practice: Moving toward a Code of Practice for Drug Development (Rejoinder)

Good Data Science Practice: Moving toward a Code of Practice for Drug Development (Rejoinder)

Application Scenarios and Practice of Data Science in the Context of Big Data

Application Scenarios and Practice of Data Science in the Context of Big Data

VLDB Scalable Data Science Category

As part of the International Conference on Very Large Data Bases (VLDB) 2021 / Proceedings of the VLDB Endowment Volume 14, a new Research Track category named Scalable Data Science (SDS) was launched [2, 6]. The goal of SDS is to attract cutting-edge and impactful real-world work in the scalable data science arena to enhance the impact and visibility of the VLDB community on data science practice, spur new technical connections, and inspire new follow-on research. The inaugural year proved to be successful, with numerous interesting papers from a wide cross section of both industry and academia, spanning several data science topics, and originating from several countries around the world. In this report, we reflect on the inaugural year of SDS with some statistics on both submissions and accepted papers, SDS invited talks, and our observations, lessons, and tips as inaugural Associate Editors for SDS. We hope this article is helpful to future authors, reviewers, and organizers of SDS, as well as other interested members of the wider database / data management community and beyond.

Disordering Datasets

The application of artificial intelligence (AI) to the behavioral health domain has led to a growing interest in the use of machine learning (ML) techniques to identify patterns in people's personal data with the goal of detecting-and even predicting-conditions such as depression, bipolar disorder, and schizophrenia. This paper investigates the data science practices and design narratives that underlie AI-mediated behavioral health through the situational analysis of three natural language processing (NLP) training datasets. Examining datasets as a sociotechnical system inextricably connected to particular social worlds, discourses, and infrastructural arrangements, we identify several misalignments between the technical project of dataset construction and benchmarking (a current focus of AI research in the behavioral health domain) and the social complexity of behavioral health. Our study contributes to a growing critical CSCW literature of AI systems by articulating the sensitizing concept ofdisordering datasets that aims to productively trouble dominant logics of AI/ML applications in behavioral health, and also support researchers and designers in reflecting on their roles and responsibilities working within this emerging and sensitive design space.