Data Science Approach Research Articles

A data science approach for multi-sensor marine observatory data monitoring cold water corals (Paragorgia arborea) in two campaigns.

Fixed underwater observatories (FUO), equipped with digital cameras and other sensors, become more commonly used to record different kinds of time series data for marine habitat monitoring. With increasing numbers of campaigns, numbers of sensors and campaign time, the volume and heterogeneity of the data, ranging from simple temperature time series to series of HD images or video call for new data science approaches to analyze the data. While some works have been published on the analysis of data from one campaign, we address the problem of analyzing time series data from two consecutive monitoring campaigns (starting late 2017 and late 2018) in the same habitat. While the data from campaigns in two separate years provide an interesting basis for marine biology research, it also presents new data science challenges, like the the marine image analysis in data form more than one campaign. In this paper, we analyze the polyp activity of two Paragorgia arborea cold water coral (CWC) colonies using FUO data collected from November 2017 to June 2018 and from December 2018 to April 2019. We successfully apply convolutional neural networks (CNN) for the segmentation and classification of the coral and the polyp activities. The result polyp activity data alone showed interesting temporal patterns with differences and similarities between the two time periods. A one month "sleeping" period in spring with almost no activity was observed in both coral colonies, but with a shift of approximately one month. A time series prediction experiment allowed us to predict the polyp activity from the non-image sensor data using recurrent neural networks (RNN). The results pave a way to a new multi-sensor monitoring strategy for Paragorgia arborea behaviour.

A data science and machine learning approach to continuous analysis of Shakespeare's plays

The availability of quantitative text analysis methods has provided new ways of analyzing literature in a manner that was not available in the pre-information era. Here we apply comprehensive machine learning analysis to the work of William Shakespeare. The analysis shows clear changes in the style of writing over time, with the most significant changes in the sentence length, frequency of adjectives and adverbs, and the sentiments expressed in the text. Applying machine learning to make a stylometric prediction of the year of the play shows a Pearson correlation of 0.71 between the actual and predicted year, indicating that Shakespeare's writing style as reflected by the quantitative measurements changed over time. Additionally, it shows that the stylometrics of some of the plays is more similar to plays written either before or after the year they were written. For instance, Romeo and Juliet is dated 1596, but is more similar in stylometrics to plays written by Shakespeare after 1600. The source code for the analysis is available for free download.

How data-driven approaches advance the search for materials relevant to energy conversion and storage

Materials discovery for proton-coupled electron transfer steps is governed by theoretical predictions using the density functional theory approximation. This method underestimates the complexity of the dynamic solid/liquid interface encountered during catalytic processes under bias. In the present perspective article, it is discussed how data-science approaches improve the analysis of the elementary steps at electrified solid/liquid interfaces by using the popular volcano concept as a tool for communication.

Local elections and the quality of financial statements in municipally owned entities: A Benford analysis

Benford’s Law has been widely used in the literature to assess data quality and reliability. This paper examines the impact of local elections on financial reporting quality in entities controlled by local governments using a data science approach. By applying Benford’s Law on financial statements published by Italian municipally owned entities operating in utility industries, we find diffuse data anomalies around election seasons. This does not automatically means that illegal manipulation or fraud are widespread in those periods, but it implies that (i) auditors need to pay particular attention to the quality of accounting data in those crucial periods and specific environments; and that (ii) voters and media have to be critical in assuming municipally owned entities’ indicators of financial performance as proxies of the administrative efficiency of incumbent politicians.

Machine learning dislocation density correlations and solute effects in Mg-based alloys

Magnesium alloys, among the lightest structural materials, represent excellent candidates for lightweight applications. However, industrial applications remain limited due to relatively low strength and ductility. Solid solution alloying has been shown to enhance Mg ductility and formability at relatively low concentrations. Zn solutes are significantly cost effective and common. However, the intrinsic mechanisms by which the addition of solutes leads to ductility improvement remain controversial. Here, by using a high throughput analysis of intragranular characteristics through data science approaches, we study the evolution of dislocation density in polycrystalline Mg and also, Mg–Zn alloys. We apply machine learning techniques in comparing electron back-scatter diffraction (EBSD) images of the samples before/after alloying and before/after deformation to extract the strain history of individual grains, and to predict the dislocation density level after alloying and after deformation. Our results are promising given that moderate predictions (coefficient of determination R^2 ranging from 0.25 to 0.32) are achieved already with a relatively small dataset (sim 5000 sub-millimeter grains).

Design Lessons from Building Deep Learning Disinformation Generation and Detection Solutions

In its essence, social media is on its way of representing the superposition of all digital representations of human concepts, ideas, believes, attitudes, and experiences. In this realm, the information is not only shared, but also {mis, dis}interpreted either unintentionally or intentionally guided by (some kind of) awareness, uncertainty, or offensive purposes. This can produce implications and consequences such as societal and political polarization, and influence or alter human behaviour and beliefs. To tackle these issues corresponding to social media manipulation mechanisms like disinformation and misinformation, a diverse palette of efforts represented by governmental and social media platforms strategies, policies, and methods plus academic and independent studies and solutions are proposed. However, such solutions are based on a technical standpoint mainly on gaming or AI-based techniques and technologies, but often only consider the defender’s perspective and address in a limited way the social perspective of this phenomenon becoming single angled. To address these issues, this research combines the defenders’ perspective with the one of the offenders by (i) building a hybrid deep learning disinformation generation and detection model and (ii) capturing and proposing a set of design recommendations that could be considered when establishing patterns, requirements, and features for building future gaming and AI-based solutions for combating social media manipulation mechanisms. This is done using the Design Science Research methodology in Data Science approach aiming at enhancing security awareness and resilience against social media manipulation.

Benford's Law for economic data reliability: The case of tourism flows in Sicily

This paper proposes a data science approach based on Benford's Law to analyse tourist flows – being tourism a relevant economic sector in Sicily. In particular, we are interested in detecting irregular patterns in the numerical data that may represent manipulations, inaccuracies or biases in the self-reported data from tourism organisations. The analysis is carried out by using monthly data for arrivals and overnight stays in hotels, B&Bs, and complementary accommodations in the seven provinces of the island from January 2016 to December 2019.We perform the analysis by employing several statistical tests and through a visual inspection of the difference between the empirical distributions and the theoretical Benford's.Conformity to Benford's distribution is mostly confirmed for the total number of overnight stays and the data considered on a yearly basis. On the contrary, we found evident deviations from Benford's Law in the empirical distribution of data broken down by nationality of tourist and accommodation type. Some comments on possible motivations for such deviations are also advanced, even though a detailed exploration of them deserves a devoted study.

A methodology for the annotation of surgical videos for supervised machine learning applications.

Surgical data science is an emerging field focused on quantitative analysis of pre-, intra-, and postoperative patient data (Maier-Hein et al. in Med Image Anal 76: 102306, 2022). Data science approaches can decompose complex procedures, train surgical novices, assess outcomes of actions, and create predictive models of surgical outcomes (Marcus et al. in Pituitary 24: 839-853, 2021; Røadsch et al. in Nat Mach Intell, 2022). Surgical videos containpowerful signals of events that may impact patient outcomes. Anecessary stepbefore the deployment of supervised machine learning methodsis the development of labels for objects andanatomy. We describe a complete method for annotating videos of transsphenoidal surgery. Endoscopic video recordings of transsphenoidal pituitary tumor removal surgeries were collected from a multicenter research collaborative. These videos were anonymized and stored in a cloud-based platform. Videos were uploaded to an online annotation platform. Annotation framework was developed based on a literature review and surgicalobservations to ensure proper understanding of the tools, anatomy, and steps present. A user guide was developed to trained annotators to ensure standardization. A fully annotated video of a transsphenoidal pituitary tumor removal surgery was produced. This annotated video included over 129,826 frames. To prevent any missing annotations, all frames were later reviewed by highly experienced annotators and a surgeon reviewer. Iterations to annotated videos allowed for the creation of an annotated video complete with labeled surgical tools, anatomy, and phases. In addition, a user guide was developed for the training of novice annotators, whichprovides information about the annotation software to ensure the production of standardized annotations. A standardized and reproducible workflow for managing surgical video data is a necessary prerequisite to surgical data science applications. We developed a standard methodology for annotatingsurgical videos that mayfacilitate the quantitative analysis of videos using machine learning applications. Future work will demonstrate the clinical relevance and impact of this workflow by developingprocess modeling andoutcome predictors.

Mechanism underlying follicular hyperproliferation and oncogenesis in hidradenitis suppurativa

Hidradenitis suppurativa (HS) is a skin disorder that causes chronic painful inflammation and hyperproliferation, often with the comorbidity of invasive keratoacanthoma (KA). Our research, employing high-resolution immunofluorescence and data science approaches together with confirmatory molecular analysis, has identified that the 5'-cap-dependent protein translation regulatory complex eIF4F is a key factor in the development of HS and is responsible for regulating follicular hyperproliferation. Specifically, eIF4F translational targets, Cyclin D1 and c-MYC, orchestrate the development of HS-associated KA. Although eIF4F and p-eIF4E are contiguous throughout HS lesions, Cyclin D1 and c-MYC have unique spatial localization and functions. The keratin-filled crater of KA is formed by nuclear c-MYC-induced differentiation of epithelial cells, whereas the co-localization of c-MYC and Cyclin D1 provides oncogenic transformation by activating RAS, PI3K, and ERK pathways. In sum, we have revealed a novel mechanism underlying HS pathogenesis of follicular hyperproliferation and the development of HS-associated invasive KA.

Current and Future Advances in Surgical Therapy for Pituitary Adenoma.

The vital physiological role of the pituitary gland, alongside its proximal critical neurovascular structures means pituitary adenomas cause significant morbidity or mortality. Whilst enormous advancements have been made in the surgical care of pituitary adenomas, treatment failure and recurrence remain challenges. To meet these clinical challenges, there has been an enormous expansion of novel medical technologies (e.g. endoscopy, advanced imaging, artificial intelligence). These innovations have the potential to benefit each step of the patient journey, and ultimately, drive improved outcomes. Earlier and more accurate diagnosis addresses this in part. Analysis of novel patient data sets, such as automated facial analysis or natural language processing of medical records holds potential in achieving an earlier diagnosis. After diagnosis, treatment decision-making and planning will benefit from radiomics and multimodal machine learning models. Surgical safety and effectiveness will be transformed by smart simulation methods for trainees. Next-generation imaging techniques and augmented reality will enhance surgical planning and intraoperative navigation. Similarly, the future armamentarium of pituitary surgeons, including advanced optical devices, smart instruments and surgical robotics, will augment the surgeon's abilities. Intraoperative support to team members will benefit from a surgical data science approach, utilising machine learning analysis of operative videos to improve patient safety and orientate team members to a common workflow. Postoperatively, early detection of individuals at risk of complications and prediction of treatment failure through neural networks of multimodal datasets will support earlier intervention, safer hospital discharge, guide follow-up and adjuvant treatment decisions. Whilst advancements in pituitary surgery hold promise to enhance the quality of care, clinicians must be the gatekeepers of technological translation, ensuring systematic assessment of risk and benefit. In doing so, the synergy between these innovations can be leveraged to drive improved outcomes for patients of the future.

Contemporary and emerging technologies for research in children's rare and interstitial lung disease.

Although recent decades have seen the identification, classification and discovery of the genetic basis of many children's interstitial and rare lung disease (chILD) disorders, detailed understanding of pathogenesis and specific therapies are still lacking for most of them. Fortunately, a revolution of technological advancements has created new opportunities to address these critical knowledge gaps. High-throughput sequencing has facilitated analysis of transcription of thousands of genes in thousands of single cells, creating tremendous breakthroughs in understanding normal and diseased cellular biology. Spatial techniques allow analysis of transcriptomes and proteomes at the subcellular level in the context of tissue architecture, in many cases even in formalin-fixed, paraffin-embedded specimens. Gene editing techniques allow creation of "humanized" animal models in a shorter time frame, for improved knowledge and preclinical therapeutic testing. Regenerative medicine approaches and bioengineering advancements facilitate the creation of patient-derived induced pluripotent stem cells and their differentiation into tissue-specific cell types which can be studied in multicellular "organoids" or "organ-on-a-chip" approaches. These technologies, singly and in combination, are already being applied to gain new biological insights into chILD disorders. The time is ripe to systematically apply these technologies to chILD, together with sophisticated data science approaches, to improve both biological understanding and disease-specific therapy.

Forecasting of residential unit’s heat demands: a comparison of machine learning techniques in a real-world case study

AbstractA large proportion of the energy consumed by private households is used for space heating and domestic hot water. In the context of the energy transition, the predominant aim is to reduce this consumption. In addition to implementing better energy standards in new buildings and refurbishing old buildings, intelligent energy management concepts can also contribute by operating heat generators according to demand based on an expected heat requirement. This requires forecasting models for heat demand to be as accurate and reliable as possible. In this paper, we present a case study of a newly built medium-sized living quarter in central Europe made up of 66 residential units from which we gathered consumption data for almost two years. Based on this data, we investigate the possibility of forecasting heat demand using a variety of time series models and offline and online machine learning (ML) techniques in a standard data science approach. We chose to analyze different modeling techniques as they can be used in different settings, where time series models require no additional data, offline ML needs a lot of data gathered up front, and online ML could be deployed from day one. A special focus lies on peak demand and outlier forecasting, as well as investigations into seasonal expert models. We also highlight the computational expense and explainability characteristics of the used models. We compare the used methods with naive models as well as each other, finding that time series models, as well as online ML, do not yield promising results. Accordingly, we will deploy one of the offline ML models in our real-world energy management system in the near future.

Data Science and Precision Oncology Nursing: Creating an Analytic Ecosystem to Support Personalized Supportive Care across the Trajectory of Illness

The authors' objective is to present an overarching framework of an analytic ecosystem using diverse data domains and data science approaches that can be used and implemented across the cancer continuum. Analytic ecosystems can improve quality practices and offer enhanced anticipatory guidance in the era of precision oncology nursing. Published scientific articles supporting the development of a novel framework with a case exemplar to provide applied examples of current barriers in data integration and use. The combination of diverse data sets and data science analytic approaches has the potential to extend precision oncology nursing research and practice. Integration of this framework can be implemented within a learning health system where models can update as new data become available across the continuum of the cancer care trajectory. To date, data science approaches have been underused in extending personalized toxicity assessments, precision supportive care, and enhancing end-of-life care practices. Nurses and nurse scientists have a unique role in the convergence of data science applications to support precision oncology across the trajectory of illness. Nurses also have specific expertise in supportive care needs that have been dramatically underrepresented in existing data science approaches thus far. They also have a role in centering the patient and family perspectives and needs as these frameworks and analytic capabilities evolve.

First Six Months of War from Ukrainian topic and sentiment analysis

Through technological advancements as well as due to societal trends and developments, social media became an active part and a catalysator of the ongoing conflicts and wars carried out in the physical environment. A direct example on this behalf are the cyber/information operations currently conducted in conjunction with the ongoing Russian-Ukrainian war. Due to such operations packaged in social media manipulation mechanisms like disinformation and misinformation using techniques such as controversies, fake news, and deep fakes, a high degree of confusion and uncertainty surrounds the events happened and users’ behaviour and beliefs. These operations also impact the civilians directly affected in the battlefield or their dear and known ones. To tackle this issue, currently limited scientific and objective effort is dedicated in this direction due to, e.g., data, strategic, and emotional implications. It is then the aim of this research to capture the main topics discussed and the feeling expressed by Ukrainian Telegram users on the ongoing Russian-Ukrainian war in 2022 using a Data Science approach by building a series of Machine Learning models based on multi-channel data collected in the first six of months of war. Accordingly, this research directly aims to contribute to efforts on understanding real discourses and dynamics involved in the ongoing conflict through direct resources, producing and sustaining social media security awareness, and building resilience to social media manipulation campaigns using AI.

Matching science to reality: how to deploy a participant-driven digital brain health platform.

Advances in digital technologies for health research enable opportunities for digital phenotyping of individuals in research and clinical settings. Beyond providing opportunities for advanced data analytics with data science and machine learning approaches, digital technologies offer solutions to several of the existing barriers in research practice that have resulted in biased samples. A participant-driven, precision brain health monitoring digital platform has been introduced to two longitudinal cohort studies, the Boston University Alzheimer's Disease Research Center (BU ADRC) and the Bogalusa Heart Study (BHS). The platform was developed with prioritization of digital data in native format, multiple OS, validity of derived metrics, feasibility and usability. A platform including nine remote technologies and three staff-guided digital assessments has been introduced in the BU ADRC population, including a multimodal smartphone application also introduced to the BHS population. Participants select which technologies they would like to use and can manipulate their personal platform and schedule over time. Participants from the BU ADRC are using an average of 5.9 technologies to date, providing strong evidence for the usability of numerous digital technologies in older adult populations. Broad phenotyping of both cohorts is ongoing, with the collection of data spanning cognitive testing, sleep, physical activity, speech, motor activity, cardiovascular health, mood, gait, balance, and more. Several challenges in digital phenotyping implementation in the BU ADRC and the BHS have arisen, and the protocol has been revised and optimized to minimize participant burden while sustaining participant contact and support. The importance of digital data in its native format, near real-time data access, passive participant engagement, and availability of technologies across OS has been supported by the pattern of participant technology use and adherence across cohorts. The precision brain health monitoring platform will be iteratively adjusted and improved over time. The pragmatic study design enables multimodal digital phenotyping of distinct clinically characterized cohorts in both rural and urban U.S. settings.

Analyzing the Role of Ukrainian and Russian Diaspora in Disinformation Campaigns

Recent digital technological developments facilitate the translation, empowerment, and extension of expressing individual and collective ideas, beliefs, and attitudes towards specific concepts and events through different social platforms. In this complex arena, social media manipulation campaigns and corresponding mechanisms like disinformation and misinformation are used through techniques like deep fakes and fake news for, e.g., altering existing information and spreading manufactured information to {targeted, diverse} audiences or producing polarization among communities and users. Nevertheless, academic and practitioner efforts to capture, control, and limit social manipulation techniques exist in the form of strategies and policies based on human intelligence, Artificial Intelligence, or a combination thereof. However, such mechanisms and consonant techniques advance in adaptivity and complexity and can reach and impact broader communities. On this behalf, and especially in conjunction with the ongoing events surrounding the ongoing war in Ukraine, increased attention and dedication is shown to both current and recent events surrounding this conflict, e.g., Crimea’s annexation and the MH17 crash in 2014. Such events characterize old battles of ongoing conflicts that could teach important lessons on understanding the role and involvement of Russian and Ukrainian diaspora communities in corresponding social manipulation discourses in social platforms like Twitter and Facebook. To tackle this, multidisciplinary research is conducted using the Design Science Research methodology following the Data Science approach building a series of Machine Learning models. Accordingly, this research aims to build and bring social awareness and resilience to both users and social media policy decision-makers on the role, involvement, and implications of diaspora digital communities in conflicts.

A Soft Power Challenge, or an Opportunity? A Big Data Analysis on Chinese Soft Power during COVID-19 Pandemic

Abstract The Chinese government's rigorous efforts to enhance its soft power have confronted a major challenge during the COVID-19 pandemic. This study aimed to look at how the Chinese soft power changed throughout the pandemic using English news articles that covered China. The research took a data science approach to investigate the contents of articles using machine-learning-based sentiment analysis and Dirichlet-Multinomial Regression (DMR) analysis. The results show a gradual downturn in overall sentiment and that the topics related to political issues made the most significant impact. Nevertheless, the major increase in referencing Chinese social media implied that the sources of Chinese soft power have been diversified throughout the pandemic. In addition, this research has aimed to engage in major debates around soft power theory. Providing a multi-disciplinary approach for analyzing soft power, this research has tackled the difficulties in the quantitative conceptualization of soft power.

Promoting Data Science in Disaster Risk Reduction: Glimpses from the Global South

There are coordinated attempts to use Data Science in disaster risk reduction (DRR) across the world. Different special interest groups are formed within the data science community to formulate different action plans and strategies to use data science in the DRR. Some examples of community-driven initiatives are the CODATA Task Group FAIR Data for Disaster Risk Research (FAIR-DRR), CODATA Task Group Linked Open Data for Global Disaster Risk Research (LODGD), and ISC Integrated Research on Disaster Risk (IRDR). This chapter highlights the data science approaches, models, frameworks, and stakeholders mapping while dealing with the data collection, analysis, and reporting for strengthening data-driven decision-making at different levels. The chapter also explores the FAIR Data Principles that aim at achieving the findable, accessible, interoperable, and reusable data resources at the institutional and national levels. FAIR Data ensures various data users increasingly engage with data related to disasters. We also discuss some case studies and success stories from the Global South, which may inspire other countries in data-driven decision-making while dealing with natural disasters. The much-discussed “Data, Information, Knowledge and Action” (DIKA) model is highlighted in this chapter. For example, the recently launched “Data-Knowledge-Action for Urban Systems” podcast series of CODATA and CEPT Research and Development Foundation (CRDF) in Ahmedabad, talks about how the DIKA model helps in creating intelligent systems for DRR using scientific and quantitative data. On a podcast in related Series, the DRR practitioners and data science researchers discussed the use of big data to predict natural disasters. This chapter highlights notable examples of how big data helped predict natural disasters across the world, particularly in the Global South.

Barriers to progress in pregnancy research: How can we break through?

Healthy pregnancies are fundamental to healthy populations, but very few therapies to improve pregnancy outcomes are available. Fundamental concepts-for example, placentation or the mechanisms that control the onset of labor-remain understudied and incompletely understood. A key issue is that research efforts must capture the complexity of the tripartite maternal-placental-fetal system, the dynamics of which change throughout gestation. Studying pregnancy disorders is complicated by the difficulty of creating maternal-placental-fetal interfaces in vitro and the uncertain relevance of animal models to human pregnancy. However, newer approaches include trophoblast organoids to model the developing placenta and integrated data-science approaches to study longer-term outcomes. These approaches provide insights into the physiology of healthy pregnancy, which is the first step to identifying therapeutic targets in pregnancy disorders.

Towards Marketing 4.0: Vision and Survey on the Role of IoT and Data Science

According to KPMG, Internet of Things (IoT) technology was among the top 10 technologies of 2019. It has been growing at a significant pace, influencing and disrupting several application domains. It is expected that by 2025, 75.44 billion devices will be connected to the Internet. These devices generate massive amounts of data which, when harnessed using the power of data science (DS) techniques and approaches such as artificial intelligence (AI) and machine learning (ML), can provide significant benefits to economy, society, and people. Examples of areas that are being disrupted are digital marketing and retail commerce services in smart cities. This paper presents a vision for Marketing 4.0 that is underpinned by disruptive digital technologies such as IoT and DS. We present an analysis of the current state of the art in IoT and DS via the three pillars of marketing: namely, people, products, and places. We propose a blueprint architecture for developing a Marketing 4.0 solution that is underpinned by IoT and DS. We conclude the paper by highlighting the open challenges that need to be addressed in order to realise the Marketing 4.0 blueprint architecture, including supporting the integration of IoT data concerning people, products, and places and using DS to make efficient and effective recommendations.